|Home | About | Journals | Submit | Contact Us | Français|
The impact of the heptavalent‐pneumococcal conjugate vaccine on the incidence of pneumococcal meningitis in Europe has not yet been assessed.
To determine whether heptavalent‐pneumococcal conjugate vaccine implementation in northern France has resulted in a decrease in the incidence of pneumococcal meningitis in children.
Multicentre retrospective cohort study from 2000 through 2005.
All paediatric departments of the 18 hospitals in northern France.
Patients <18 years of age, admitted for laboratory‐confirmed pneumococcal meningitis during the study period, were included.
Data were collected from medical files and the microbiological laboratories of each hospital and compared with the regional hospital discharge codes, using a capture–recapture method.
The study assessed and compared global and age‐related incidence rates of pneumococcal meningitis in 2001 (pre‐vaccine era) and 2005.
77 cases were found through the capture–recapture method. The incidence rate of pneumococcal meningitis varied from 1.65/100000 children <18 years in 2001 to 0.80/100000 children in 2005 (53% reduction, 95% CI 31 to 74; p=0.08). This has so far been significant only for children <2 years of age (8.9/100000 in 2001 to 1.8/100000 in 2005; 82% reduction, 95% CI 52 to 95; p=0.03).
A decline in pneumococcal meningitis has been observed in infants since heptavalent‐pneumococcal conjugate vaccination began in our area.
In the United States, Streptococcus pneumoniae has been considered to be the principal pathogen for bacterial meningitis (47%) since Haemophilus influenzae type b vaccination became widespread during the 1990s and before the implementation of vaccination with the heptavalent‐pneumococcal conjugate vaccine (PCV7, Prevenar).1 In Western Europe, the mean incidence rate of pneumococcal meningitis has averaged 8.7 cases/100000 in children <2 years old with incidence rates varying from 3.8 to 14.6/100000 between countries.2 Between 2001 and 2004, the French Bacterial Meningitis Surveillance Network reported that S pneumoniae caused 42% of all cases of bacterial meningitis in children, 70% of these occurring in children <2 years old.3 The case‐fatality rate for this disease is estimated at 8–12% in children and has not dramatically changed for 20 years despite progress in diagnosis and treatment.1,4 Sequelae occur in 20–35% of cases and include deafness, motor deficits, learning disorders linked with concentration disorders, and memory problems.5,6
PCV7, first approved in the US in 1999, targets the seven serotypes involved most frequently in the invasive pneumococcal diseases of young children.7,8 Serotypes 6B, 9V, 14, 18C and 23F, all present in this vaccine, account for most cases of pneumococcal meningitis today.9 PCV7 received marketing authorisation in Europe in February 2001, was available in France in April 2001 and was recommended in March 2002 for children with a disease at high risk of invasive pneumococcal infections (immunosuppression, sickle cell disease, etc),10 and children aged 2–24 months with risk factors for pneumococcal infection (ie, children cared for more than 4 h/week with more than two other children, children with breast‐feeding duration <2 months, children with at least two siblings), criteria which covered between 79 and 89% of children <2 years of age.11 The vaccination schedule uses a four‐dose regimen, at 2, 3 and 4 months of age and a booster dose during the second year of life. The impact of the PCV7 on the incidence of meningitis and other invasive pneumococcal diseases has been clearly demonstrated in North America and Australia,12,13,14,15,16 whose vaccination schedules are different from those in France.
The aim of this study was to determine whether PCV7 implementation in a large area of northern France affected the incidence of pneumococcal meningitis in children.
We conducted a descriptive, retrospective, multicentre cohort study covering a 6‐year period (2000–2005) and involving all paediatric departments of the 18 hospitals in the region (1 million children <18 years old). All units agreed to participate in this study. In each centre, the local coordinating investigator was a paediatrician belonging to the Hospital Network for Evaluating the Management of Common Childhood Diseases. Each patient <18 years old and admitted for pneumococcal meningitis during the study period was eligible for inclusion. Pneumococcal meningitis was defined as the acute onset of meningitis (cerebrospinal fluid (CSF) white blood cell (WBC) count 7/mm3)17,18 and documented S pneumoniae infection in either CSF (direct examination, culture, PCR or latex agglutination) or blood (culture or PCR). The study excluded patients with traumatic lumbar puncture (CSF red blood cells >10000/mm3),19 purpura fulminans, bacterial contamination or pre‐treated meningitis with negative microbiological tests for which bacterial meningitis could not be ascertained.
The primary identification of cases was performed by the local coordinating investigator, by cross‐checking medical files with laboratory records. Demographic data (age, sex, date of admission), clinical features (seizure, petechial rash, haemodynamic failure, altered consciousness), laboratory results (WBC and neutrophil counts in blood and CSF, C reactive protein in blood, protein and glucose concentrations in CSF, CSF Gram staining, blood and CSF culture, soluble antigens, PCR in blood or CSF, and serotyping of microorganisms if any) were recorded for each patient. These data were extracted from medical files in each paediatric unit and from the bacterial meningitis registry of the microbiological laboratory in each hospital. They were reported on an anonymous standardised data form by the investigating paediatrician at each hospital and sent to the principal investigators (FD, IM), who arranged for the data to be entered into a database. This database was compared with the regional hospital discharge codes (which provided age, sex, dates of hospital stay, diagnosis, origin of the patients, and discharge diagnosis codes), using the ICD‐10 codes (G0.01) as the primary or associated discharge diagnosis, to assess the corrected incidence of our case ascertainment procedure. For this type of retrospective study, neither approval from an ethics committee nor consent from patients or their parents was required in France.
The PCV7 manufacturer furnished us with a list of PCV7 doses sold in northern France during the study period. The National Reference Centre provided us with the pneumococcal serotyping results for children in the cohort study to determine whether PCV7 diffusion had modified serotypes.
Statistical analyses was carried out using Epi Info 6.04 software (Centers for Disease Control and Prevention, Atlanta, GA, USA). Analysis began by describing the characteristics of patients with pneumococcal meningitis. We used a capture–recapture method,20 and demographic data from the national institute for statistics and economic studies (www.insee.fr) to determine the number and corrected incidence rates of pneumococcal meningitis in children <18 years old between 2000 and 2005. Incidence rates in 2001 (the year PCV7 was approved in Europe) and 2005 (the last year of data collection) were compared with a χ2 test. The percentage change in this rate was calculated together with its 95% confidence interval (CI). The relationship between the number of vaccine doses sold and the estimated number of pneumococcal meningitis cases, as well as the annual age distribution of identified pneumococcal meningitis cases, were analysed. p Values <0.05 were considered to indicate statistical significance.
During the study period, 65 children, identified through medical and laboratory records in the 18 hospitals with paediatric departments, were diagnosed with pneumococcal meningitis. Data from patients with this disease admitted in 2000 could not be collected at one centre. Cross‐checking with discharge codes (56 cases found) revealed 73 notified patients (table 11).). The case ascertainment rate of medical and laboratory records was 89%, while the discharge codes missed 23% of the diagnoses. The children's mean age was 31 months (range 28 days–14.9 years), with 63% <2 years old, and the male–female ratio was 0.9. Admission to the intensive care unit was necessary in 27% of cases, with haemodynamic failure in 5% and neurological failure in 51%. CSF Gram staining was positive for 76% of patients. The mean CSF WBC count was 2359 cells/mm3 (range 7–15700), with a mean CSF neutrophil count of 83% (range 18–100). Mean CSF glucose and protein concentrations were 34 and 187 mg/dl, respectively. Mean CRP was 173 mg/l (range 2–429). Overall, 3% of the children died and 42% had neurological sequelae.
The capture–recapture method estimated that 77 children overall had pneumococcal meningitis during the study period in northern France (table 11).). Using these estimates for each year from 2001 to 2005, we calculated that the incidence of pneumococcal meningitis decreased from 17 cases (1.65/100000 children <18 years old) to 8 cases/year (0.80/100000), which corresponded to a 53% reduction (95% CI 31 to 74; p=0.08).
As fig 11 shows, the estimated number of pneumococcal meningitis cases decreased as vaccine distribution increased in northern France. The reduction in cases affected only children younger than 2 years of age (fig 22).). Incidence rose briefly from 8.9/100000 children <2 years old in 2001 to 10.5/100000 in 2002, and subsequently fell to 1.8/100000 in 2005 (82% reduction; 95% CI 52 to 95; p=0.03). No such decrease was observed for patients aged 2–5 years and >5 years old. Less than two thirds of the strains were sent for serotyping to the national reference centre from 2001 to 2004, whereas 88% were serotyped in 2005. Three of the seven serotyped strains were related to strains included in the vaccine in 2005, five of five in 2004, and five of seven in 2003. Of the 13 patients with pneumococcal meningitis caused by strains included in the vaccine since 2003, 11 were not vaccinated (10 could have been vaccinated and one belonged to the old unvaccinated cohort of patients), one had received two doses and one had received three doses.
Our multicentre study of pneumococcal meningitis found 77 children through the capture–recapture method with this disease in the study area during the study period (2000–2005). Incidence of the disease among children <18 years old decreased between 2001 and 2005, with a 53% reduction during the study period, but only among those less than 2 years was the decrease, from 8.9 to 1.8/100000 between 2001 and 2005, statistically significant (p=0.03). This is the first study in Europe to demonstrate that the incidence of pneumococcal meningitis decreased significantly between 2001 and 2005 among a large population of children, that is, following the implementation of PCV7.
This vaccine has had a significant impact in North America and Australia over the past few years: the incidence of invasive pneumococcal diseases, including meningitis, has fallen by 56–69% in children less than 2 years of age.14,15,16,21,22,23 PCV7 coverage, however, is substantially higher in the US (68% of children had at least three injections in 2003)24 than in France (41% of children were vaccinated in 2004 with three doses).11 Moreover, the seven pneumococcal serotypes contained in this conjugate vaccine were responsible for 83% of the cases of invasive pneumococcal diseases in young children in the USA,7 but only 68% of the blood isolates and 61% of CSF isolates in France.25 Given the differences in policies, vaccination schedules,10 vaccine coverage rates and serotype distributions in different countries, vaccination effects evidently cannot be expected to be similar between them. For these reasons, studies of the effects of this vaccine elsewhere are essential.
Although our study covered a large area of northern France (1million children <18 years old), rather few cases of childhood pneumococcal meningitis were found, but this is consistent with previously reported incidence rates.2 However, incidence changed significantly after PCV7 implementation began. The retrospective design of our study may have contributed to a selection bias. For example, no data from the year 2000 were available for one centre, which explains why results from that year were considered for only part of the analyses. This did not, however, interfere with the main results of the study. Moreover, the quality of our data collection was controlled by cross‐checking with laboratory records before reporting back to the study coordinators and comparing our database with regional hospital discharge codes. Unlike the existing national surveillance programs,26 the capture–recapture method allowed us to determine the corrected incidence of pneumococcal meningitis for children <18 years of age in northern France. A study of a network of French laboratories recently found a 31% reduction in cases of pneumococcal bacteraemia in children <2 years old, including pneumococcal meningitis.27 However, variations in reporting rates, estimated at 61–79%, may have contributed to this result.4,26,27 Our study did not include suspected bacterial meningitides without positive bacteriology, since inclusion criteria required laboratory confirmation. Patients' characteristics were similar to those usually reported.3 Data concerning pneumococcal serotypes are less informative, since less than two thirds of strains were sent to the national reference centre for serotyping (except in 2005). Only a few strains were serotyped per year, with inconclusive results. Data for the year 2005 require further confirmation.
Although the PCV7 coverage rate is steadily increasing in France, it was still low in June 2004, when 41% of children had been vaccinated; only 27% had had three doses by the age of 6 months.11 In 2005, it was estimated from the number of doses sold that 63% of children <1 year old had been vaccinated nationwide.28 In May 2006, indications in France were simplified to cover every child <2 years old, which should increase the number of children vaccinated.28
Our results are also important because pneumococcal resistance most often affects children <2 years old: 71% of pneumococcal strains are not susceptible to penicillin and 18% are resistant in French children with invasive infections.25 In the United States, a significant decrease in pneumococcal resistance has been noted since widespread pneumococcal vaccination began.14,22 A reduction in antibiotic‐resistant pneumococcal strains has also been observed in France, with multiple drug resistance dropping from 15.4% in 2001 to 6.7% in 2004 among nasopharyngeal serotype samples from children aged 6–24 months with otitis.29 Nasopharyngeal carriage of non‐vaccine serotypes with a normal or reduced sensitivity to penicillin has been noted in young children,30 and recent studies report an increase in invasive pneumococcal diseases from new serotypes.31,32,33 Further surveillance and the introduction of new conjugate vaccines are needed to maximise the benefits of the conjugate pneumococcal vaccine.
The reduction in the incidence rate of pneumococcal meningitis between 2001 and 2005 in northern France must be confirmed by further local surveillance and by the national observatory of bacterial meningitis. Widespread extension of this PCV7 coverage rate will facilitate further reduction, unless new serotypes replace those currently in the vaccine.
We are grateful to the paediatricians and microbiologists in northern France participating in the Hospital Network for Evaluating the Management of Common Childhood Diseases who contributed to the present study: Dr Akitani, Seclin Hospital; Dr Audry‐Degardin, Valenciennes Hospital; Dr Blondiaux, Cambrai Hospital; Dr Chenaud, Tourcoing Hospital; Dr Delepoulle, Dunkerque Hospital; Dr Devouge, Arras Hospital; Dr El Kohen, Lille–Saint‐Vincent‐de‐Paul Hospital; Dr Evrard, Boulogne Hospital; Dr Glowacki, Armentières Hospital; Dr Gnansoungou, Maubeuge Hospital; Dr Guilly, Saint‐Omer Hospital; Dr Louf, Montreuil‐sur‐mer Hospital; Dr Martinet, Béthune Hospital; Dr Pierre, Roubaix Hospital; Dr Racoussot, Douai Hospital; Dr Renault, Calais Hospital; and Dr Vanlaeys, Lens Hospital.
Many thanks to Dr Varon from the Centre National de Référence du Pneumocoque, Georges Pompidou European Hospital, Paris, for providing pneumococcal serotypes for patients of this cohort study.
CSF - cerebrospinal fluid
WBC - white blood cell
Competing interests: None.