- Email: [email protected]
Elimination of Haemophilus influenzae type b (Hib) disease from The Gambia after the introduction of routine immunisation with a Hib conjugate vaccine: a prospective study
Articles
Elimination of Haemophilus influenzae type b (Hib) disease from The Gambia after the introduction of routine immunisation with a Hib conjugate vaccine: a prospective study Richard A Adegbola, Ousman Secka, George Lahai, Nellie Lloyd-Evans, Alpha Njie, Stanley Usen, Claire Oluwalana, Stephen Obaro, Martin Weber, Tumani Corrah, Kim Mulholland, Keith McAdam, Brian Greenwood, Paul J M Milligan
Summary Lancet 2005; 366: 144–50 Published online June 29, 2005 DOI:10.1016/S0140-6736(05) 66788-8
Background Routine immunisation of infants in The Gambia with a Haemophilus influenzae type b (Hib) polysaccharide-tetanus toxoid conjugate vaccine began in May, 1997. We investigated the effectiveness of the vaccine when delivered through the expanded programme on immunisation and the effect of national immunisation on incidence of Hib disease.
See Comment page 101 Medical Research Council Laboratories, The Gambia (R Adegbola FRCPath, O Secka MPhil, G Lahai BSc, S Usen FWACP, C Oluwalana MBBS, S Obaro FRCPCH, M Weber MD, Prof K McAdam FRCP, T Corrah FRCP); National Health Laboratory (N Lloyd-Evans MSc) and Department of State for Health, Banjul, The Gambia (A Njie MPH); Royal Children’s Hospital, Victoria, Australia (Prof K Mulholland FRACP); and Department of Infectious Diseases, London School of Hygiene and Tropical Medicine, London, UK (Prof B Greenwood FRCP, P Milligan PhD) Correspondence to: Dr Richard A Adegbola, MRC Laboratories, P O Box 273, Banjul, The Gambia [email protected]
144
Methods Surveillance for Hib disease was maintained in the western half of The Gambia using standard methods with an emphasis on meningitis. We estimated vaccine efficacy using the case control method, and vaccine coverage and population denominators for incidence rates using a cluster sample survey. Prevalence of Hib carriage in a sample of 1–2-year old children attending health centres for vaccination was ascertained with oropharyngeal swabs plated onto antiserum agar. Findings Between May, 1997, and April, 2002, a total of 5984 children were examined for possible Hib infections. 49 children had Hib disease, 36 of whom had meningitis. The annual incidence rates of Hib meningitis before any use of the vaccine (1990–93) dropped from over 200 per 100 000 children aged younger than 1 year to none per 100 000 in 2002, and from 60 to no cases per 100 000 in children younger than 5 years. The prevalence of Hib carriage decreased from 12% to 0·25% (p0·0001). Two doses of vaccine were needed for direct protection from Hib disease (vaccine efficacy 94%, 95% CI 62–99). Since most children received a protective dose after the age of greatest disease risk, indirect effects were important in reducing disease incidence. Interpretation The Gambian Hib immunisation programme reduced the occurrence of Hib disease despite irregular vaccine supply. The effect of the programme in The Gambia has important implications for the introduction of the vaccine into routine immunisation programmes of other developing countries.
Introduction Before the widespread use of conjugate vaccines for Haemophilus influenzae type b (Hib), this organism was the most common cause of severe invasive childhood infections in developed countries.1 The introduction of routine vaccination in developed countries has led to substantial declines in the incidence of Hib disease.2,3 However, routine use of these vaccines in less developed countries has been delayed until recently4,5 because there was little information available about the burden of Hib disease and because the high price of conjugate vaccines made them unaffordable.6 Studies undertaken in The Gambia before the Hib conjugate vaccine was investigated showed that the epidemiology of Hib disease was different from that seen in more developed countries. Overall incidence was higher, infections occurred in children of younger ages, and pneumonia was more prevalent and related to a worse outcome in The Gambia than in more developed countries.7,8 The case fatality rate of Hib meningitis was 30%,9 sequelae were common, and less than half of patients recovered fully from Hib disease.10 However,
epiglottitis was not reported in The Gambia or any other developing countries.11 A successful trial of a Hib conjugate vaccine was undertaken in 42 000 infants in The Gambia during 1993 to 1995.12 The vaccine used was a Hib polysaccharide-tetanus toxoid conjugate (PRP-T; Act-Hib, supplied by Pasteur Mérieux, Lyon, France). The vaccine showed 95% efficacy against cultureconfirmed invasive Hib disease, 21% efficacy against radiologically defined pneumonia with definite alveolar consolidation, and 60% protection against carriage in the second year of life.12,13 However, reports of vaccine efficacy under optimum trial conditions do not guarantee that a vaccine will be effective in routine use. This possible situation is especially relevant in developing countries where suboptimum storage and transport conditions can reduce vaccine efficacy. Indirect effects (herd immunity) can make vaccination more effective than suggested by an individually randomised efficacy trial. Additionally, national immunisation could change the epidemiology of Hib disease in developing countries, leading to the emergence of epiglottitis. www.thelancet.com Vol 366 July 9, 2005
Articles
Routine immunisation of Gambian infants with PRP-T vaccine was introduced on May 1, 1997, via the Gambia’s national expanded programme on immunisation (EPI). The Gambia’s EPI is one of the most successful immunisation programmes in Africa. Staff at health centres throughout the country administer standard EPI vaccines (BCG, diphtheria-tetanus-pertussis [DTP], polio, measles, hepatitis B, yellow fever). DTP vaccinations are scheduled to be given at 2 months, 3 months, and 4 months of age although, in practice, they are usually given later. Vaccine coverage is assessed regularly by use of the cluster survey technique, recommended by WHO.14 The coverage of DTP3 has been more than 85% since 1990. The vaccine was supplied as part of a donation agreement between Pasteur Mérieux (now Sanofi Pasteur MSD), the vaccine manufacturers, and the Gambia Government, mediated by UNICEF. Since 2002, supply of Hib vaccine has continued with support from the Global Alliance for Vaccines and Immunization (GAVI). We aimed to determine the effectiveness of PRP-T vaccine in preventing Hib meningitis among Gambian children when delivered through The Gambia’s EPI since 1997. We also sought to estimate the contribution of indirect effects, associated with reduced throat carriage of Hib in the second year of life, to the reduction of disease incidence. Finally, we endeavoured to document any unusual or late presentations of Hib disease occurring in children who received vaccination in early infancy during the efficacy trial.
Methods Study population Surveillance for Hib disease was restricted to the western region of The Gambia. This region includes rural, urban, and rapidly growing periurban areas, with a total population in 2003 of 836 000—60% of the total population of the country (Ndow ASM, Department of Central Statistics, Banjul, The Gambia, unpublished, 2004). We enrolled children in the present study if they had been investigated as possible cases of Hib disease and if they were younger than 6 years. Children from the previous randomised trial12 were eligible for inclusion in our study without any age restrictions. Those who presented at hospital with a rapid respiratory rate, fever, and signs of meningitis, and those with clinical signs of pneumonia or septicaemia, were investigated. When a case of invasive Hib disease had been confirmed, vaccination status, date of birth, and other details were obtained from the child’s health card and the mother or guardian was interviewed with the same questions as those used for the controls (described later). In 1999, 2000, and 2001, the prevalence of Hib carriage in children aged 1–2 years who had attended health centres for routine immunisation was established. In each year, we selected a systematic www.thelancet.com Vol 366 July 9, 2005
sample of 500 children at Sibanor, a rural health centre, and 500 at Serrekunda, an urban health centre, during October to January, and an oropharyngeal swab was obtained from every child.
Surveillance procedures The incidence of Hib disease in the western region of The Gambia was first surveyed between August, 1985, and July, 1987.9 Since 1990, continuous surveillance for Hib disease has been maintained in the same area with standard methods.8 Investigations for possible Hib disease are undertaken at three hospitals. Two hospitals, the Royal Victoria Teaching Hospital, Banjul, and the Medical Research Council Hospital, Fajara, are in urban areas and one, WEC Mission Sibanor Hospital, is in a rural area. The catchment population of Sibanor Hospital is much smaller than the other hospitals. The Royal Victoria Teaching Hospital is the main government referral hospital in the western region. Surveillance for invasive Hib disease started in May, 1997, and continued until April 30, 2002, and focused on the detection of meningitis, the most readily diagnosed form of invasive Hib disease. To estimate accurate denominators for incidence rates we determined the number of children in single age categories, in 1993, from the national census (which formed our baseline census data for calculating growth rate from 1993 to 2000) using Sprague’s method, after smoothing the age structure with Arriaga’s method (Sonko S, unpublished data). The number of children aged 1–2 years in 2000 was ascertained by use of a cluster sample survey with probability sampling. The number of children aged 0–1 years and 0–5 years in each year of the study was then interpolated on the basis of a constant exponential growth rate chosen so that the projection coincided with the survey estimates in 2000 (age-specific estimates were not available from the 2003 census at the time of writing).
Case definitions A patient was diagnosed as having bacterial meningitis if their leucocyte count in their cerebrospinal fluid was more than 10 cells per L and if bacteria were isolated from their cerebrospinal fluid. The presence of clinical or radiological pneumonia and isolation of Hib from blood or lung aspirate were categorised as Hib pneumonia. Patients with no focus of infection from whom bacteria were isolated from blood cultures were classified as having septicaemia. To assist health personnel in the recognition of epiglottitis, a video of patients with this syndrome was kindly provided by Professor Frank Shann (Royal Children’s Hospital, Melbourne, Australia).
Vaccine delivery and vaccination coverage Vaccination is delivered through 20 health centres and through monthly outreach clinics. The number of 145
Articles
doses delivered was recorded in each health centre using tally sheets; these data were checked against stocks and collated every 3 months by study staff. We did a cluster sample survey at the end of 2000 to estimate vaccination coverage using a probability sampling method designed to produce estimates of the proportion vaccinated with three doses with a margin of error of 6%. We also estimated the total number of vaccinated and unvaccinated children. 60 clusters (enumeration areas of the 1993 census) were selected with probability proportional to estimated population size, and 828 children aged 12–23 months were surveyed.14 Vaccination status was determined from the child health card or from the mother’s recall if the health card was missing.
Vaccine efficacy We did a case control study between May 1, 1997, and Dec 31, 2001, to estimate direct effects of vaccination (vaccine efficacy). We included all children with confirmed meningitis, pneumonia, or septicaemia due to Hib. Soon after a case was detected, we identified ten controls for each case, born within 2 weeks of the patient and living in the same locality at the time they became sick. A fieldworker located the compound where the case was living when they became sick, and then moved two compounds away and looked for matched controls. At each compound the number of children younger than 3 years without a health card was recorded. When a child was found who was born within 2 weeks of the case, vaccination details were copied from the health card and the mother or carer was interviewed and a questionnaire was completed. A subset of cases and controls were visited again by a 12 000
Hib trial
National immunisation with Hib vaccine
To estimate the relative contribution of direct and indirect effects of vaccination we used the age distribution of Hib meningitis cases recorded during 1990–93, before use of Hib vaccine in The Gambia, and the distribution of age at vaccination, from our survey in 2000, to calculate the proportion of cases that could be prevented by direct effects alone, with the assumption that there was no shift in age distribution of disease. If “A” represents the age of the disease case, and “V” the age 2 weeks after vaccination with the second dose (the point when the dose is considered to be effective), then the probability P(VA) is the probability of receiving a protective dose early enough to prevent Hib disease. Multiplying this probability by the estimated vaccine efficacy VE, P(VA)VE, gives an estimate of the expected proportion of disease cases that could be prevented by the direct effects of vaccination. This calculation can be compared with the actual reduction in disease incidence to see whether direct effects alone could explain the observed reduction in disease incidence.
8000 1·0 0·9 6000
0·8
4000
2000
0
Jan
–1
99 5 Jan –1 99 6 Jan –1 99 7 Jan –1 99 8 Jan –1 99 9 Jan –2 00 0 Jan –2 00 1 Jan –2 00 2
4
3 Jan
–1
99
2
99 –1 Jan
99 Jan
–1
99
1
0 99
–1 Jan
–1 Jan
0·7 0·6 0·5 0·4 0·3
DPT/Hib dose 1
0·2
DPT/Hib dose 2
0·1
DPT/Hib dose 3
0
Figure 1: Number of doses of Hib conjugate vaccine delivered per month in the western region of The Gambia during the efficacy trial and in the first 5 years of routine delivery in the EPI programme
146
Indirect effects of vaccination
Vaccination coverage
Doses of Hib vaccine delivered per month
10 000
second fieldworker to check the vaccination details. Information was gathered from cases and controls about variables thought to be related to the risk of disease, including distance from health centre, overcrowding as measured by the number of children younger than 2 years in the compound, and mother’s education, for inclusion as potential confounders in the analysis of vaccine efficacy. Date of birth, vaccination status, and dates of vaccination were obtained from the health card. Health cards were seen for all the cases; we excluded around 3% of possibly eligible controls because they did not have a health card. For the first and second doses of vaccine the dose was not thought to be effective until 14 days after vaccination. For the third dose, the corresponding interval was 21 days.12 We estimated vaccine efficacy using conditional logistic regression (Stata version 7, Statacorp, Texas).
0
2
4
6
8 10 12 Age (months)
14
16
18
20
Figure 2: Coverage with one, two, or three doses of DPT/Hib vaccine, in 2000 Proportion vaccinated by the specified age in a sample of children aged 12–23 months in December, 2000.
www.thelancet.com Vol 366 July 9, 2005
Articles
A
A
50
400 Hib trial
National immunisation with Hib
Hib trial
National immunisation with Hib
350 40
Meningitis cases
300 250
30
200 20
150 100
10
50 0 B
0
10
B 100
Meningitis cases
90 80 70 5 60 50 40 30 0 0
12
24
36
48
60
Age in months
20 10
Bacteriology Hib and other bacterial pathogens were isolated and identified with standard methods. Additionally, Hib isolates were confirmed by standard slide agglutination with type-specific antisera (Murex Diagnostics, Dartford, UK) and capsular genotyping by PCR amplification of capsule-specific sequences, as described previously.12 The joint ethics committee of the Gambian government and the MRC approved the study and a local steering committee was formed to oversee the study’s day-to-day progress.
Role of the funding source The sponsors 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 Introduction of Hib vaccine into the routine EPI immunisation schedule was implemented on May 1, 1997. Health centre staff were trained to record DTP and Hib vaccinations on the standard health card; children born after May 1, 1997, were issued with a modified www.thelancet.com Vol 366 July 9, 2005
2 00
1 00
Jan –2
0 00
Jan –2
9 99
–2 Jan
8 99
Jan –1
7 99
–1 Jan
6 99
Jan –1
5 99
Jan –1
4 99
Jan –1
3 99
–1 Jan
2 99
Jan –1
1 99
Jan –1
99 Jan –1
Jan –1
0
0
Figure 3: Age distribution of meningitis cases A: 1990–93, before any use of Hib vaccine. B: 1997–2002, after introduction of national immunisation with Hib vaccine.
Figure 4: Incidence of Hib meningitis per 100 000 per year, in the western region of The Gambia A: children younger than 1 year. B: children younger than 5 years. Dotted lines represent pointwise 90% likelihood-based confidence limits.
health card that explicitly outlined Hib vaccination. Routine immunisation with Hib and DTP was interrupted from time to time because of a meningitis vaccination campaign (July to August, 1997) and irregular supplies of vaccine (figure 1). 828 children in 60 clusters in The Gambia were surveyed in December 2000, of whom 95% had a health card providing an accurate record of immunisation status. Coverage with three doses of DTP/Hib was 68% (95% CI 62–75%), with one dose was 94% (92–97), and with two doses was 84% (79–88; figure 2). The estimated number of children aged 12–23 months in the study region was 22 960 (95% CI 18 706–27 214), so that an additional 12 398 doses of vaccine would have been needed per year to bring coverage of three doses up to 100%. The median age at vaccination was 3·4 months (dose 1), 6·5 months (dose 2), and 8 months (dose 3). A total of 5984 children were investigated as possible cases of Hib infection; 3019 (50%) were from the Medical Research Council Hospital, 2547 (43%) were from the Royal Victoria Teaching Hospital, and 418 (7%) 147
Articles
Number of cases
0 1 2 3
Meningitis Pneumonia
Septicaemia
23 10 2 1
1 2 0 1
5 1 0 0
Vaccine efficacy against all Hib disease* (95% CI)
Vaccine efficacy for meningitis cases only* (95% CI)
38% (–58 to 75) p=0·317 94% (62 to 99) p=0·003 94% (62 to 99) p=0·003
35% (–84 to 77) p=0·421 93% (42 to 99) p=0·013 96% (47 to 100) p=0·015
*Adjusted for the effects of covariates; unadjusted estimates for one, two, and three doses were 48%, 93%, and 95% for all Hib disease, respectively, and 36%, 88%, and 94% for meningitis only.
Table: Vaccine efficacy against Hib disease after one, two, and three doses of vaccine
were from Sibanor. 2244 (38%) children were younger than 1 year of age. The number of suspected cases of Hib disease that were investigated was similar in each year of the study; 1075 (18%) in 1997, 1120 (19%) in 1998, 846 (14%) in 1999, 1327 (22%) in 2000, and 1303 (22%) in 2001. 49 children with Hib disease were identified; 36 (74%) had meningitis, eight (16%) had pneumonia, and five (10%) had septicaemia. 38 (78%) were younger than 12 months of age. 28 (57%) had not received Hib vaccine, 17 (35%) had received only one dose, two (4%) had received two doses, and two (4%) had received three doses. Both children who were fully immunised died; one had Hib septicaemia and was infected with HIV-1, the other child, who had Hib meningitis, was admitted to hospital in a rural area and died before further investigation. After consent was obtained, this child’s parents and siblings were screened for HIV infection, but they all tested negative. There were no cases of Hib disease among children who had been enrolled in the previous randomised trial12 and no child with epiglottitis was reported. Nine (25%) of 36 children with meningitis died, 24 (67%) were discharged fully recovered, two (6%) were discharged with sequelae, and one was lost to follow-up. 0·25 Hib trial
National immunisation with Hib
Prevalence of Hib carriage
0·20
0·15 Rural health centre Urban health centre
0·10
0·05
02 20
01 Jan
20
00 Jan
20
99 Jan
8 99
19 Jan
97
99 6
19
Jan 1
Jan
95 19
Jan 1
19
94 Jan
19
93 Jan
92 Jan
1 99
19 Jan
19 Jan
Jan 1
90
0
Figure 5: Prevalence of Hib carriage among children attending a rural and an urban health centre for routine immunisation before and after the start of national immunisation with Hib vaccine Vertical lines indicate 95% CI.
148
31 of the 36 (86%) children with meningitis were 12 months or younger and 18 (50%) were 6 months or younger. The age distribution of meningitis cases was very similar to that recorded in 1990–93, before the start of the vaccine trial (figure 3). 22 of the cases (61%) were identified during the first year (1997/98) of routine Hib vaccination. The numbers of cases during the subsequent 4 years were five, six, none, and three, respectively. Thus, the numbers of children with Hib meningitis decreased from 22 in 1997/98 to three in 2001, by contrast with the 36–40 cases per year that had been identified from the same community before the introduction of Hib vaccine. There were 37 cases of Hib meningitis in children younger than 1 year of age during a 3-year period before The Gambian trial (1990/93), nine cases during the 2-year period of the efficacy trial (1993/95), 14 cases in the year before the introduction of routine Hib vaccination (1996/97), and six cases during the present study period. We calculated the incidence of Hib meningitis in 6-month periods from January, 1990, to January, 2002 (figure 4). The annual incidence of Hib meningitis in children younger than 1 year old declined rapidly from over 200 cases per 100 000 children before any use of Hib in The Gambia, to 0 per 100 000 in the last year of this study. Similarly, the incidence rates of Hib meningitis decreased from 60 to 0 cases per 100 000 in children younger than 5 years old (figure 4). 49 cases of Hib disease were identified during the study. Controls were not obtained for two cases who lived outside The Gambia and 11 cases were not included in the matched analysis because the case and all matched controls were unvaccinated. The point estimates of efficacy of one, two, and three doses of vaccine were much the same when the analysis was restricted to meningitis cases only or to all cases of Hib disease. The estimated vaccine efficacy against meningitis after two doses was 93% and there was no apparent benefit from an additional dose (table). Vaccine efficacy against Hib disease after one dose was 38%, but confidence intervals were wide (–58% to 75%). 7 (19%) of the cases of meningitis detected during the 5-year study were younger than 3 months of age, too young to have been vaccinated. This finding compares with 11 cases (39%) younger than 3 months of age in the 2·5 years of surveillance before the vaccine trial. The median age of all Hib meningitis cases in the prevaccine period was 7 months (IQR 4·9–9·3), and the median age at which the second dose of Hib vaccine was received after introduction of national immunisation was 6·5 months (IQR 4·6–9·0). The estimate of P(VA) was 0·44 and vaccine efficacy of two doses was 93%, so that a reduction in disease incidence of only 41% would be expected due to direct effects alone. Restricting the calculation to disease cases younger than 1 year of age, we obtain a figure of 35% predicted reduction in incidence among children younger than 1 year due to direct effects. Therefore, the greater part of the observed reduction in disease incidence was probably due to indirect effects. www.thelancet.com Vol 366 July 9, 2005
Articles
15 children had Hib disease caused by non-type b serotypes, six had meningitis, eight had septicaemia, and one child had pneumonia. Five of the Hib isolated were of serotype a, one was of serotype f, and the remaining nine were not serotypable. Five of the eight children with septicaemia also had other disorders; four had malnutrition and one had measles. This number of cases of non-type b Hib disease was similar to the number expected on the basis of those identified before the introduction of Hib vaccine.9 165 children with pneumococcal disease were identified during the study period; 44 had meningitis, 80 had pneumonia, and 41 had septicaemia. The number of children detected with pneumococcal disease was closely similar each year of the surveillance period; 35 (21%) for each of 1997 and 1999, 31 (19%) for 1998, 34 (21%) for 2000, and 19 (12%) for 2001. A comparison of the carriage rates with those obtained before the start of the efficacy trial in 1993 shows a progressive decline in Hib carriage (figure 5). In the period before the trial started (1992–93), 236 (12%, 95% CI 10–13) of 2000 children sampled were positive for Hib, compared with 5 (0%, 0·08–0·6) of 2000 sampled in 2000–01.
Discussion After the introduction of routine Hib immunisation, the annual incidence rate of meningitis in infants in The Gambia dropped sharply from over 200 per 100 000 to 21 per 100 000 in the first 2 years of vaccination.5 The incidence declined further to 0 per 100 000 in the fifth year after the introduction of routine immunisation. This reduction was achieved despite difficulties with a regular supply of vaccine. No cases of Hib disease have been identified in the western region since 2002. Surveillance for bacterial diseases was undertaken in two other divisions, located in more rural areas of The Gambia, during a pneumococcal vaccine efficacy trial between August, 2000, and April, 2004.15 The trial results showed a closely similar trend for Hib disease; a total of 17 437 children aged 40–364 days were enrolled over 3 years from a total population of 380 000. 11 cases of Hib were reported over the course of the trial. Four of these were children aged 1–11 months who had meningitis; two were detected in 2001 and two in 2002 (Cutts F, personal communication). A cluster sample survey of the whole country, undertaken in 2000, showed that vaccine coverage of DTP/Hib3 was between 83% and 96% in these more rural areas of the country compared with 68% in the study area. The efficacy of two doses of Hib vaccine against Hib disease was 94%; efficacy after a single dose was only 38% and not statistically significant. Our estimate of vaccine efficacy after a single dose has a wide confidence interval; however, our results indicate that two doses are needed for protection and lend support to findings of several other studies.16,17 Most children received their www.thelancet.com Vol 366 July 9, 2005
second dose too late to be protected directly. The observed reduction in incidence therefore shows, to a large extent, an indirect effect of the vaccine, probably through an effect on carriage. There seemed to be little added benefit from a third dose. Only two of the children with Hib disease had received three doses of vaccine, and one of these had HIV-1 infection. No children were reported to have epiglottitis despite a special effort to identify cases. Routine use of Hib vaccination has not resulted in the emergence of unexpected forms of Hib disease in older children (2 years) in The Gambia. During the previous, individually randomised Gambian vaccine trial,12 Hib vaccination gave 60% protective efficacy against carriage of Hib among vaccinated 1–2-year olds. After nationwide vaccination, carriage of Hib has virtually been eliminated from both vaccinated and unvaccinated children indicating that vaccination has reduced transmission of Hib in The Gambia to an almost negligible level. The effect on carriage was recorded in both urban and rural areas, with carriage being highest in rural areas. Although several studies have suggested that an increase in non-Hib invasive disease18,19 or pneumococcal disease20,21 occurs after the introduction of Hib vaccination, there was no evidence of this in our study. Our results lend support to other reports that have not shown any increase in non-Hib invasive disease,22 pneumococcal disease, or meningococcal disease23 after the introduction of Hib vaccine. There have been reports of increasing numbers of Hib cases in the UK where booster immunisation with Hib vaccine has not been used; 7 years after an initial rapid decrease in the rate of invasive Hib disease after the introduction of childhood immunisation with Hib vaccine24,25 the introduction of booster immunisation is being considered. Thus, sustained surveillance will be needed in The Gambia, where booster immunisation is not given, to ensure that the effect of national immunisation with Hib vaccine is sustained. This goal is being achieved through a new regional programme of surveillance for Hib meningitis with support from WHO AFRO regional office. The remarkable effect of the Hib vaccination programme on incidence of Hib disease in The Gambia has important implications for the introduction of the vaccine into routine immunisation programmes of other developing countries. This successful outcome was achieved despite an erratic supply of vaccine and indicates that virtual elimination of Hib disease can be accomplished in a developing country setting with an EPI programme that is effective but not perfect. We hope that the results of this study will encourage countries in sub-Saharan Africa that have not yet adopted routine Hib vaccination to do so and encourage those that have done so to sustain this highly effective way of preventing death and severe disability related to Hib disease in young children. 149
Articles
Contributors R A Adegbola coordinated the study and, together with study paediatricians S Usen, C Oluwalana, S Obaro, and M Weber, was responsible for enrolment and ascertainment of study cases. R A Adegbola, N Lloyd-Evans, and O Secka were responsible for processing and interpretation of all bacteriology specimens and for the carriage studies. P Milligan was responsible for design and organisation of the case control study and the survey of vaccination coverage, for collection of data on vaccine delivery, and for data analysis. A Njie, K McAdam, and T Corrah contributed to study coordination. G Lahai was responsible for data entry and management. B Greenwood and K Mulholland made contributions to study concept and protocol development. R Adegbola and P Milligan wrote the paper with input from all authors.
9
Conflict of interest statement We declare that we have no conflict of interest.
13
Acknowledgments The WHO Department of Vaccines and Biologicals and the MRC provided financial support for this study. We thank the staff of the Sibanor WEC Mission Hospital and paediatric and nursing staff of Royal Victoria Teaching Hospital and MRC Laboratories Hospital for their enthusiastic support; Peter Smith and Shabbar Jaffar for advice; Usman Ikumapayi for his technical assistance at the RVTH bacteriology laboratory; the field staff and data entry clerks who worked on this project; and the Department of Central Statistics, Banjul, The Gambia Divisional Health Team, and EPI staff who implemented the introduction of routine Hib vaccination. References 1 Cochi SL, O’Mara D, Preblud SR. Progress in Haemophilus type b polysaccharide vaccine use in the United States. Pediatrics 1988; 81: 166–68. 2 Peltola H, Kilpi T, Anttila M. Rapid disappearance of Haemophilus influenzae type b meningitis after routine childhood immunisation with conjugate vaccines. Lancet 1992; 340: 592–94. 3 Adams WG, Deaver KA, Cochi SL, et al. Decline of childhood Haemophilus influenzae type b (Hib) disease in the Hib vaccine era. JAMA 1993; 269: 221–26. 4 Lagos R, Levine OS, Avendano A, Horwitz I, Levine MM. The introduction of routine Haemophilus influenzae type b conjugate vaccine in Chile: a framework for evaluating new vaccines in newly industrialized countries. Pediatr Infect Dis J 1998; 17 (suppl 9): S139–48. 5 Adegbola RA, Usen SO, Weber M, et al. Haemophilus influenzae type b meningitis in The Gambia after introduction of a conjugate vaccine. Lancet 1999; 354: 1091–92. 6 Wenger JD, DiFabio J-L, Landaverde JM, Levine OS, Gaafar T. Introduction of Hib conjugate vaccines in the non-industrialised world: experience in four “newly adopting” countries. Vaccine 2000; 18: 736–42. 7 Bijlmer HA. World-wide epidemiology of Haemophilus influenzae meningitis; industrialized versus non-industrialized countries. Vaccine 1991; 9 (suppl): S5–9. 8 Adegbola RA, Mulholland EK, Falade AG, et al. Haemophilus influenzae type b disease in the Western Region of the Gambia: background surveillance for a vaccine efficacy trial. Ann Trop Paediatr 1996; 16: 103–11.
150
10
11
12
14
15
16
17
18
19
20 21 22 23
24 25
Bijlmer HA, van Alphen L, Greenwood BM, et al. The epidemiology of Haemophilus influenzae meningitis in children under five years of age in The Gambia, West Africa. J Infect Dis 1990; 161: 1210–15. Goetghebuer T, West TE, Wermenbol V, et al. Outcome of meningitis caused by Streptococcus pneumoniae and Haemophilus influenzae type b in children in The Gambia. Trop Med Int Health 2000; 5: 207–13. Greenwood BM. Epidemiology of acute lower respiratory tract infections, especially those due to Haemophilus influenzae type b, in The Gambia, West Africa. J Infect Dis 1992; 165: S26–28. Mulholland K, Hilton S, Adegbola R, et al. Randomised trial of Haemophilus influenzae type-b tetanus protein conjugate vaccine for prevention of pneumonia and meningitis in Gambian infants. Lancet 1997; 349: 1191–97. Adegbola RA, Mulholland EK, Secka O, Jaffar S, Greenwood BM. Vaccination with a Haemophilus influenzae type b conjugate vaccine reduces oropharyngeal carriage of H influenzae type b among Gambian children. J Infect Dis 1998; 177: 1758–61. Milligan PJM, Njie A, Bennett S. A comparison of two cluster sampling methods for health surveys in developing countries. Int J Epidemiol 2004; 33: 469–76. Cutts FT, Zaman SMA, Enwere G, et al. Efficacy of nine-valent pneumococcal conjugate vaccine against pneumonia and invasive pneumococcal disease in The Gambia: randomised, double-blind, placebo-controlled trial. Lancet 2005; 365: 1139–46. Decker MD, Edwards KM, Bradley R, Palmer P. Comparative trial in infants of four conjugate Haemophilus inlfuenzae type b vaccines. J Pediatr 1992; 120: 184–89. Bulkow LR, Wainwright RB, Letson GW, Chang SJ, Ward JI. Comparative immunogenicity of four Heamophilus influenzae type b conjugate vaccines in Alaska Native infants. Pediatr Infect Dis J 1993; 12: 484–92. Waggoner-Fountain L, Hendley O, Cody EJ, Perriello VA, Donowitz LG. The emergence of Haemophilus influenzae types e and f as significant pathogens. Clin Infect Dis 1995; 21: 1322–24. Ward JI. Editorial response: invasive infections due to Haemophilus influenzae serotype F (Hif) Is Hif an emerging pathogen? Clin Infect Dis 1996; 22: 1077–78. Baer M, Vuento R, Vesikari T. Increase in bacteremic pneumococcal infections in children. Lancet 1995; 345: 661. Gieseche J, Fredlund H. Increase in pneumococcal bacteremia in Sweden. Lancet 1997; 349: 670–99. Jordens JZ, Slack MPE. Haemophilus influenzae: then and now. Eur J Clin Microbiol Infect Dis 1995; 14: 935–48. Dagan R, Fraser D, Roitman M, et al. Effectiveness of a nationwide infant immunization program against Haemophilus influenzae type b. Vaccine 1999; 17: 134–41. Garner D, Weston V. Effectiveness of vaccination for Haemophilus influenzae type b. Lancet 2003; 361: 395–96. 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.
www.thelancet.com Vol 366 July 9, 2005
Elimination of Haemophilus influenzae type b (Hib) disease from The Gambia after the introduction of routine immunisation with a Hib conjugate vaccine: a prospective study Richard A Adegbola, Ousman Secka, George Lahai, Nellie Lloyd-Evans, Alpha Njie, Stanley Usen, Claire Oluwalana, Stephen Obaro, Martin Weber, Tumani Corrah, Kim Mulholland, Keith McAdam, Brian Greenwood, Paul J M Milligan
Summary Lancet 2005; 366: 144–50 Published online June 29, 2005 DOI:10.1016/S0140-6736(05) 66788-8
Background Routine immunisation of infants in The Gambia with a Haemophilus influenzae type b (Hib) polysaccharide-tetanus toxoid conjugate vaccine began in May, 1997. We investigated the effectiveness of the vaccine when delivered through the expanded programme on immunisation and the effect of national immunisation on incidence of Hib disease.
See Comment page 101 Medical Research Council Laboratories, The Gambia (R Adegbola FRCPath, O Secka MPhil, G Lahai BSc, S Usen FWACP, C Oluwalana MBBS, S Obaro FRCPCH, M Weber MD, Prof K McAdam FRCP, T Corrah FRCP); National Health Laboratory (N Lloyd-Evans MSc) and Department of State for Health, Banjul, The Gambia (A Njie MPH); Royal Children’s Hospital, Victoria, Australia (Prof K Mulholland FRACP); and Department of Infectious Diseases, London School of Hygiene and Tropical Medicine, London, UK (Prof B Greenwood FRCP, P Milligan PhD) Correspondence to: Dr Richard A Adegbola, MRC Laboratories, P O Box 273, Banjul, The Gambia [email protected]
144
Methods Surveillance for Hib disease was maintained in the western half of The Gambia using standard methods with an emphasis on meningitis. We estimated vaccine efficacy using the case control method, and vaccine coverage and population denominators for incidence rates using a cluster sample survey. Prevalence of Hib carriage in a sample of 1–2-year old children attending health centres for vaccination was ascertained with oropharyngeal swabs plated onto antiserum agar. Findings Between May, 1997, and April, 2002, a total of 5984 children were examined for possible Hib infections. 49 children had Hib disease, 36 of whom had meningitis. The annual incidence rates of Hib meningitis before any use of the vaccine (1990–93) dropped from over 200 per 100 000 children aged younger than 1 year to none per 100 000 in 2002, and from 60 to no cases per 100 000 in children younger than 5 years. The prevalence of Hib carriage decreased from 12% to 0·25% (p0·0001). Two doses of vaccine were needed for direct protection from Hib disease (vaccine efficacy 94%, 95% CI 62–99). Since most children received a protective dose after the age of greatest disease risk, indirect effects were important in reducing disease incidence. Interpretation The Gambian Hib immunisation programme reduced the occurrence of Hib disease despite irregular vaccine supply. The effect of the programme in The Gambia has important implications for the introduction of the vaccine into routine immunisation programmes of other developing countries.
Introduction Before the widespread use of conjugate vaccines for Haemophilus influenzae type b (Hib), this organism was the most common cause of severe invasive childhood infections in developed countries.1 The introduction of routine vaccination in developed countries has led to substantial declines in the incidence of Hib disease.2,3 However, routine use of these vaccines in less developed countries has been delayed until recently4,5 because there was little information available about the burden of Hib disease and because the high price of conjugate vaccines made them unaffordable.6 Studies undertaken in The Gambia before the Hib conjugate vaccine was investigated showed that the epidemiology of Hib disease was different from that seen in more developed countries. Overall incidence was higher, infections occurred in children of younger ages, and pneumonia was more prevalent and related to a worse outcome in The Gambia than in more developed countries.7,8 The case fatality rate of Hib meningitis was 30%,9 sequelae were common, and less than half of patients recovered fully from Hib disease.10 However,
epiglottitis was not reported in The Gambia or any other developing countries.11 A successful trial of a Hib conjugate vaccine was undertaken in 42 000 infants in The Gambia during 1993 to 1995.12 The vaccine used was a Hib polysaccharide-tetanus toxoid conjugate (PRP-T; Act-Hib, supplied by Pasteur Mérieux, Lyon, France). The vaccine showed 95% efficacy against cultureconfirmed invasive Hib disease, 21% efficacy against radiologically defined pneumonia with definite alveolar consolidation, and 60% protection against carriage in the second year of life.12,13 However, reports of vaccine efficacy under optimum trial conditions do not guarantee that a vaccine will be effective in routine use. This possible situation is especially relevant in developing countries where suboptimum storage and transport conditions can reduce vaccine efficacy. Indirect effects (herd immunity) can make vaccination more effective than suggested by an individually randomised efficacy trial. Additionally, national immunisation could change the epidemiology of Hib disease in developing countries, leading to the emergence of epiglottitis. www.thelancet.com Vol 366 July 9, 2005
Articles
Routine immunisation of Gambian infants with PRP-T vaccine was introduced on May 1, 1997, via the Gambia’s national expanded programme on immunisation (EPI). The Gambia’s EPI is one of the most successful immunisation programmes in Africa. Staff at health centres throughout the country administer standard EPI vaccines (BCG, diphtheria-tetanus-pertussis [DTP], polio, measles, hepatitis B, yellow fever). DTP vaccinations are scheduled to be given at 2 months, 3 months, and 4 months of age although, in practice, they are usually given later. Vaccine coverage is assessed regularly by use of the cluster survey technique, recommended by WHO.14 The coverage of DTP3 has been more than 85% since 1990. The vaccine was supplied as part of a donation agreement between Pasteur Mérieux (now Sanofi Pasteur MSD), the vaccine manufacturers, and the Gambia Government, mediated by UNICEF. Since 2002, supply of Hib vaccine has continued with support from the Global Alliance for Vaccines and Immunization (GAVI). We aimed to determine the effectiveness of PRP-T vaccine in preventing Hib meningitis among Gambian children when delivered through The Gambia’s EPI since 1997. We also sought to estimate the contribution of indirect effects, associated with reduced throat carriage of Hib in the second year of life, to the reduction of disease incidence. Finally, we endeavoured to document any unusual or late presentations of Hib disease occurring in children who received vaccination in early infancy during the efficacy trial.
Methods Study population Surveillance for Hib disease was restricted to the western region of The Gambia. This region includes rural, urban, and rapidly growing periurban areas, with a total population in 2003 of 836 000—60% of the total population of the country (Ndow ASM, Department of Central Statistics, Banjul, The Gambia, unpublished, 2004). We enrolled children in the present study if they had been investigated as possible cases of Hib disease and if they were younger than 6 years. Children from the previous randomised trial12 were eligible for inclusion in our study without any age restrictions. Those who presented at hospital with a rapid respiratory rate, fever, and signs of meningitis, and those with clinical signs of pneumonia or septicaemia, were investigated. When a case of invasive Hib disease had been confirmed, vaccination status, date of birth, and other details were obtained from the child’s health card and the mother or guardian was interviewed with the same questions as those used for the controls (described later). In 1999, 2000, and 2001, the prevalence of Hib carriage in children aged 1–2 years who had attended health centres for routine immunisation was established. In each year, we selected a systematic www.thelancet.com Vol 366 July 9, 2005
sample of 500 children at Sibanor, a rural health centre, and 500 at Serrekunda, an urban health centre, during October to January, and an oropharyngeal swab was obtained from every child.
Surveillance procedures The incidence of Hib disease in the western region of The Gambia was first surveyed between August, 1985, and July, 1987.9 Since 1990, continuous surveillance for Hib disease has been maintained in the same area with standard methods.8 Investigations for possible Hib disease are undertaken at three hospitals. Two hospitals, the Royal Victoria Teaching Hospital, Banjul, and the Medical Research Council Hospital, Fajara, are in urban areas and one, WEC Mission Sibanor Hospital, is in a rural area. The catchment population of Sibanor Hospital is much smaller than the other hospitals. The Royal Victoria Teaching Hospital is the main government referral hospital in the western region. Surveillance for invasive Hib disease started in May, 1997, and continued until April 30, 2002, and focused on the detection of meningitis, the most readily diagnosed form of invasive Hib disease. To estimate accurate denominators for incidence rates we determined the number of children in single age categories, in 1993, from the national census (which formed our baseline census data for calculating growth rate from 1993 to 2000) using Sprague’s method, after smoothing the age structure with Arriaga’s method (Sonko S, unpublished data). The number of children aged 1–2 years in 2000 was ascertained by use of a cluster sample survey with probability sampling. The number of children aged 0–1 years and 0–5 years in each year of the study was then interpolated on the basis of a constant exponential growth rate chosen so that the projection coincided with the survey estimates in 2000 (age-specific estimates were not available from the 2003 census at the time of writing).
Case definitions A patient was diagnosed as having bacterial meningitis if their leucocyte count in their cerebrospinal fluid was more than 10 cells per L and if bacteria were isolated from their cerebrospinal fluid. The presence of clinical or radiological pneumonia and isolation of Hib from blood or lung aspirate were categorised as Hib pneumonia. Patients with no focus of infection from whom bacteria were isolated from blood cultures were classified as having septicaemia. To assist health personnel in the recognition of epiglottitis, a video of patients with this syndrome was kindly provided by Professor Frank Shann (Royal Children’s Hospital, Melbourne, Australia).
Vaccine delivery and vaccination coverage Vaccination is delivered through 20 health centres and through monthly outreach clinics. The number of 145
Articles
doses delivered was recorded in each health centre using tally sheets; these data were checked against stocks and collated every 3 months by study staff. We did a cluster sample survey at the end of 2000 to estimate vaccination coverage using a probability sampling method designed to produce estimates of the proportion vaccinated with three doses with a margin of error of 6%. We also estimated the total number of vaccinated and unvaccinated children. 60 clusters (enumeration areas of the 1993 census) were selected with probability proportional to estimated population size, and 828 children aged 12–23 months were surveyed.14 Vaccination status was determined from the child health card or from the mother’s recall if the health card was missing.
Vaccine efficacy We did a case control study between May 1, 1997, and Dec 31, 2001, to estimate direct effects of vaccination (vaccine efficacy). We included all children with confirmed meningitis, pneumonia, or septicaemia due to Hib. Soon after a case was detected, we identified ten controls for each case, born within 2 weeks of the patient and living in the same locality at the time they became sick. A fieldworker located the compound where the case was living when they became sick, and then moved two compounds away and looked for matched controls. At each compound the number of children younger than 3 years without a health card was recorded. When a child was found who was born within 2 weeks of the case, vaccination details were copied from the health card and the mother or carer was interviewed and a questionnaire was completed. A subset of cases and controls were visited again by a 12 000
Hib trial
National immunisation with Hib vaccine
To estimate the relative contribution of direct and indirect effects of vaccination we used the age distribution of Hib meningitis cases recorded during 1990–93, before use of Hib vaccine in The Gambia, and the distribution of age at vaccination, from our survey in 2000, to calculate the proportion of cases that could be prevented by direct effects alone, with the assumption that there was no shift in age distribution of disease. If “A” represents the age of the disease case, and “V” the age 2 weeks after vaccination with the second dose (the point when the dose is considered to be effective), then the probability P(VA) is the probability of receiving a protective dose early enough to prevent Hib disease. Multiplying this probability by the estimated vaccine efficacy VE, P(VA)VE, gives an estimate of the expected proportion of disease cases that could be prevented by the direct effects of vaccination. This calculation can be compared with the actual reduction in disease incidence to see whether direct effects alone could explain the observed reduction in disease incidence.
8000 1·0 0·9 6000
0·8
4000
2000
0
Jan
–1
99 5 Jan –1 99 6 Jan –1 99 7 Jan –1 99 8 Jan –1 99 9 Jan –2 00 0 Jan –2 00 1 Jan –2 00 2
4
3 Jan
–1
99
2
99 –1 Jan
99 Jan
–1
99
1
0 99
–1 Jan
–1 Jan
0·7 0·6 0·5 0·4 0·3
DPT/Hib dose 1
0·2
DPT/Hib dose 2
0·1
DPT/Hib dose 3
0
Figure 1: Number of doses of Hib conjugate vaccine delivered per month in the western region of The Gambia during the efficacy trial and in the first 5 years of routine delivery in the EPI programme
146
Indirect effects of vaccination
Vaccination coverage
Doses of Hib vaccine delivered per month
10 000
second fieldworker to check the vaccination details. Information was gathered from cases and controls about variables thought to be related to the risk of disease, including distance from health centre, overcrowding as measured by the number of children younger than 2 years in the compound, and mother’s education, for inclusion as potential confounders in the analysis of vaccine efficacy. Date of birth, vaccination status, and dates of vaccination were obtained from the health card. Health cards were seen for all the cases; we excluded around 3% of possibly eligible controls because they did not have a health card. For the first and second doses of vaccine the dose was not thought to be effective until 14 days after vaccination. For the third dose, the corresponding interval was 21 days.12 We estimated vaccine efficacy using conditional logistic regression (Stata version 7, Statacorp, Texas).
0
2
4
6
8 10 12 Age (months)
14
16
18
20
Figure 2: Coverage with one, two, or three doses of DPT/Hib vaccine, in 2000 Proportion vaccinated by the specified age in a sample of children aged 12–23 months in December, 2000.
www.thelancet.com Vol 366 July 9, 2005
Articles
A
A
50
400 Hib trial
National immunisation with Hib
Hib trial
National immunisation with Hib
350 40
Meningitis cases
300 250
30
200 20
150 100
10
50 0 B
0
10
B 100
Meningitis cases
90 80 70 5 60 50 40 30 0 0
12
24
36
48
60
Age in months
20 10
Bacteriology Hib and other bacterial pathogens were isolated and identified with standard methods. Additionally, Hib isolates were confirmed by standard slide agglutination with type-specific antisera (Murex Diagnostics, Dartford, UK) and capsular genotyping by PCR amplification of capsule-specific sequences, as described previously.12 The joint ethics committee of the Gambian government and the MRC approved the study and a local steering committee was formed to oversee the study’s day-to-day progress.
Role of the funding source The sponsors 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 Introduction of Hib vaccine into the routine EPI immunisation schedule was implemented on May 1, 1997. Health centre staff were trained to record DTP and Hib vaccinations on the standard health card; children born after May 1, 1997, were issued with a modified www.thelancet.com Vol 366 July 9, 2005
2 00
1 00
Jan –2
0 00
Jan –2
9 99
–2 Jan
8 99
Jan –1
7 99
–1 Jan
6 99
Jan –1
5 99
Jan –1
4 99
Jan –1
3 99
–1 Jan
2 99
Jan –1
1 99
Jan –1
99 Jan –1
Jan –1
0
0
Figure 3: Age distribution of meningitis cases A: 1990–93, before any use of Hib vaccine. B: 1997–2002, after introduction of national immunisation with Hib vaccine.
Figure 4: Incidence of Hib meningitis per 100 000 per year, in the western region of The Gambia A: children younger than 1 year. B: children younger than 5 years. Dotted lines represent pointwise 90% likelihood-based confidence limits.
health card that explicitly outlined Hib vaccination. Routine immunisation with Hib and DTP was interrupted from time to time because of a meningitis vaccination campaign (July to August, 1997) and irregular supplies of vaccine (figure 1). 828 children in 60 clusters in The Gambia were surveyed in December 2000, of whom 95% had a health card providing an accurate record of immunisation status. Coverage with three doses of DTP/Hib was 68% (95% CI 62–75%), with one dose was 94% (92–97), and with two doses was 84% (79–88; figure 2). The estimated number of children aged 12–23 months in the study region was 22 960 (95% CI 18 706–27 214), so that an additional 12 398 doses of vaccine would have been needed per year to bring coverage of three doses up to 100%. The median age at vaccination was 3·4 months (dose 1), 6·5 months (dose 2), and 8 months (dose 3). A total of 5984 children were investigated as possible cases of Hib infection; 3019 (50%) were from the Medical Research Council Hospital, 2547 (43%) were from the Royal Victoria Teaching Hospital, and 418 (7%) 147
Articles
Number of cases
0 1 2 3
Meningitis Pneumonia
Septicaemia
23 10 2 1
1 2 0 1
5 1 0 0
Vaccine efficacy against all Hib disease* (95% CI)
Vaccine efficacy for meningitis cases only* (95% CI)
38% (–58 to 75) p=0·317 94% (62 to 99) p=0·003 94% (62 to 99) p=0·003
35% (–84 to 77) p=0·421 93% (42 to 99) p=0·013 96% (47 to 100) p=0·015
*Adjusted for the effects of covariates; unadjusted estimates for one, two, and three doses were 48%, 93%, and 95% for all Hib disease, respectively, and 36%, 88%, and 94% for meningitis only.
Table: Vaccine efficacy against Hib disease after one, two, and three doses of vaccine
were from Sibanor. 2244 (38%) children were younger than 1 year of age. The number of suspected cases of Hib disease that were investigated was similar in each year of the study; 1075 (18%) in 1997, 1120 (19%) in 1998, 846 (14%) in 1999, 1327 (22%) in 2000, and 1303 (22%) in 2001. 49 children with Hib disease were identified; 36 (74%) had meningitis, eight (16%) had pneumonia, and five (10%) had septicaemia. 38 (78%) were younger than 12 months of age. 28 (57%) had not received Hib vaccine, 17 (35%) had received only one dose, two (4%) had received two doses, and two (4%) had received three doses. Both children who were fully immunised died; one had Hib septicaemia and was infected with HIV-1, the other child, who had Hib meningitis, was admitted to hospital in a rural area and died before further investigation. After consent was obtained, this child’s parents and siblings were screened for HIV infection, but they all tested negative. There were no cases of Hib disease among children who had been enrolled in the previous randomised trial12 and no child with epiglottitis was reported. Nine (25%) of 36 children with meningitis died, 24 (67%) were discharged fully recovered, two (6%) were discharged with sequelae, and one was lost to follow-up. 0·25 Hib trial
National immunisation with Hib
Prevalence of Hib carriage
0·20
0·15 Rural health centre Urban health centre
0·10
0·05
02 20
01 Jan
20
00 Jan
20
99 Jan
8 99
19 Jan
97
99 6
19
Jan 1
Jan
95 19
Jan 1
19
94 Jan
19
93 Jan
92 Jan
1 99
19 Jan
19 Jan
Jan 1
90
0
Figure 5: Prevalence of Hib carriage among children attending a rural and an urban health centre for routine immunisation before and after the start of national immunisation with Hib vaccine Vertical lines indicate 95% CI.
148
31 of the 36 (86%) children with meningitis were 12 months or younger and 18 (50%) were 6 months or younger. The age distribution of meningitis cases was very similar to that recorded in 1990–93, before the start of the vaccine trial (figure 3). 22 of the cases (61%) were identified during the first year (1997/98) of routine Hib vaccination. The numbers of cases during the subsequent 4 years were five, six, none, and three, respectively. Thus, the numbers of children with Hib meningitis decreased from 22 in 1997/98 to three in 2001, by contrast with the 36–40 cases per year that had been identified from the same community before the introduction of Hib vaccine. There were 37 cases of Hib meningitis in children younger than 1 year of age during a 3-year period before The Gambian trial (1990/93), nine cases during the 2-year period of the efficacy trial (1993/95), 14 cases in the year before the introduction of routine Hib vaccination (1996/97), and six cases during the present study period. We calculated the incidence of Hib meningitis in 6-month periods from January, 1990, to January, 2002 (figure 4). The annual incidence of Hib meningitis in children younger than 1 year old declined rapidly from over 200 cases per 100 000 children before any use of Hib in The Gambia, to 0 per 100 000 in the last year of this study. Similarly, the incidence rates of Hib meningitis decreased from 60 to 0 cases per 100 000 in children younger than 5 years old (figure 4). 49 cases of Hib disease were identified during the study. Controls were not obtained for two cases who lived outside The Gambia and 11 cases were not included in the matched analysis because the case and all matched controls were unvaccinated. The point estimates of efficacy of one, two, and three doses of vaccine were much the same when the analysis was restricted to meningitis cases only or to all cases of Hib disease. The estimated vaccine efficacy against meningitis after two doses was 93% and there was no apparent benefit from an additional dose (table). Vaccine efficacy against Hib disease after one dose was 38%, but confidence intervals were wide (–58% to 75%). 7 (19%) of the cases of meningitis detected during the 5-year study were younger than 3 months of age, too young to have been vaccinated. This finding compares with 11 cases (39%) younger than 3 months of age in the 2·5 years of surveillance before the vaccine trial. The median age of all Hib meningitis cases in the prevaccine period was 7 months (IQR 4·9–9·3), and the median age at which the second dose of Hib vaccine was received after introduction of national immunisation was 6·5 months (IQR 4·6–9·0). The estimate of P(VA) was 0·44 and vaccine efficacy of two doses was 93%, so that a reduction in disease incidence of only 41% would be expected due to direct effects alone. Restricting the calculation to disease cases younger than 1 year of age, we obtain a figure of 35% predicted reduction in incidence among children younger than 1 year due to direct effects. Therefore, the greater part of the observed reduction in disease incidence was probably due to indirect effects. www.thelancet.com Vol 366 July 9, 2005
Articles
15 children had Hib disease caused by non-type b serotypes, six had meningitis, eight had septicaemia, and one child had pneumonia. Five of the Hib isolated were of serotype a, one was of serotype f, and the remaining nine were not serotypable. Five of the eight children with septicaemia also had other disorders; four had malnutrition and one had measles. This number of cases of non-type b Hib disease was similar to the number expected on the basis of those identified before the introduction of Hib vaccine.9 165 children with pneumococcal disease were identified during the study period; 44 had meningitis, 80 had pneumonia, and 41 had septicaemia. The number of children detected with pneumococcal disease was closely similar each year of the surveillance period; 35 (21%) for each of 1997 and 1999, 31 (19%) for 1998, 34 (21%) for 2000, and 19 (12%) for 2001. A comparison of the carriage rates with those obtained before the start of the efficacy trial in 1993 shows a progressive decline in Hib carriage (figure 5). In the period before the trial started (1992–93), 236 (12%, 95% CI 10–13) of 2000 children sampled were positive for Hib, compared with 5 (0%, 0·08–0·6) of 2000 sampled in 2000–01.
Discussion After the introduction of routine Hib immunisation, the annual incidence rate of meningitis in infants in The Gambia dropped sharply from over 200 per 100 000 to 21 per 100 000 in the first 2 years of vaccination.5 The incidence declined further to 0 per 100 000 in the fifth year after the introduction of routine immunisation. This reduction was achieved despite difficulties with a regular supply of vaccine. No cases of Hib disease have been identified in the western region since 2002. Surveillance for bacterial diseases was undertaken in two other divisions, located in more rural areas of The Gambia, during a pneumococcal vaccine efficacy trial between August, 2000, and April, 2004.15 The trial results showed a closely similar trend for Hib disease; a total of 17 437 children aged 40–364 days were enrolled over 3 years from a total population of 380 000. 11 cases of Hib were reported over the course of the trial. Four of these were children aged 1–11 months who had meningitis; two were detected in 2001 and two in 2002 (Cutts F, personal communication). A cluster sample survey of the whole country, undertaken in 2000, showed that vaccine coverage of DTP/Hib3 was between 83% and 96% in these more rural areas of the country compared with 68% in the study area. The efficacy of two doses of Hib vaccine against Hib disease was 94%; efficacy after a single dose was only 38% and not statistically significant. Our estimate of vaccine efficacy after a single dose has a wide confidence interval; however, our results indicate that two doses are needed for protection and lend support to findings of several other studies.16,17 Most children received their www.thelancet.com Vol 366 July 9, 2005
second dose too late to be protected directly. The observed reduction in incidence therefore shows, to a large extent, an indirect effect of the vaccine, probably through an effect on carriage. There seemed to be little added benefit from a third dose. Only two of the children with Hib disease had received three doses of vaccine, and one of these had HIV-1 infection. No children were reported to have epiglottitis despite a special effort to identify cases. Routine use of Hib vaccination has not resulted in the emergence of unexpected forms of Hib disease in older children (2 years) in The Gambia. During the previous, individually randomised Gambian vaccine trial,12 Hib vaccination gave 60% protective efficacy against carriage of Hib among vaccinated 1–2-year olds. After nationwide vaccination, carriage of Hib has virtually been eliminated from both vaccinated and unvaccinated children indicating that vaccination has reduced transmission of Hib in The Gambia to an almost negligible level. The effect on carriage was recorded in both urban and rural areas, with carriage being highest in rural areas. Although several studies have suggested that an increase in non-Hib invasive disease18,19 or pneumococcal disease20,21 occurs after the introduction of Hib vaccination, there was no evidence of this in our study. Our results lend support to other reports that have not shown any increase in non-Hib invasive disease,22 pneumococcal disease, or meningococcal disease23 after the introduction of Hib vaccine. There have been reports of increasing numbers of Hib cases in the UK where booster immunisation with Hib vaccine has not been used; 7 years after an initial rapid decrease in the rate of invasive Hib disease after the introduction of childhood immunisation with Hib vaccine24,25 the introduction of booster immunisation is being considered. Thus, sustained surveillance will be needed in The Gambia, where booster immunisation is not given, to ensure that the effect of national immunisation with Hib vaccine is sustained. This goal is being achieved through a new regional programme of surveillance for Hib meningitis with support from WHO AFRO regional office. The remarkable effect of the Hib vaccination programme on incidence of Hib disease in The Gambia has important implications for the introduction of the vaccine into routine immunisation programmes of other developing countries. This successful outcome was achieved despite an erratic supply of vaccine and indicates that virtual elimination of Hib disease can be accomplished in a developing country setting with an EPI programme that is effective but not perfect. We hope that the results of this study will encourage countries in sub-Saharan Africa that have not yet adopted routine Hib vaccination to do so and encourage those that have done so to sustain this highly effective way of preventing death and severe disability related to Hib disease in young children. 149
Articles
Contributors R A Adegbola coordinated the study and, together with study paediatricians S Usen, C Oluwalana, S Obaro, and M Weber, was responsible for enrolment and ascertainment of study cases. R A Adegbola, N Lloyd-Evans, and O Secka were responsible for processing and interpretation of all bacteriology specimens and for the carriage studies. P Milligan was responsible for design and organisation of the case control study and the survey of vaccination coverage, for collection of data on vaccine delivery, and for data analysis. A Njie, K McAdam, and T Corrah contributed to study coordination. G Lahai was responsible for data entry and management. B Greenwood and K Mulholland made contributions to study concept and protocol development. R Adegbola and P Milligan wrote the paper with input from all authors.
9
Conflict of interest statement We declare that we have no conflict of interest.
13
Acknowledgments The WHO Department of Vaccines and Biologicals and the MRC provided financial support for this study. We thank the staff of the Sibanor WEC Mission Hospital and paediatric and nursing staff of Royal Victoria Teaching Hospital and MRC Laboratories Hospital for their enthusiastic support; Peter Smith and Shabbar Jaffar for advice; Usman Ikumapayi for his technical assistance at the RVTH bacteriology laboratory; the field staff and data entry clerks who worked on this project; and the Department of Central Statistics, Banjul, The Gambia Divisional Health Team, and EPI staff who implemented the introduction of routine Hib vaccination. References 1 Cochi SL, O’Mara D, Preblud SR. Progress in Haemophilus type b polysaccharide vaccine use in the United States. Pediatrics 1988; 81: 166–68. 2 Peltola H, Kilpi T, Anttila M. Rapid disappearance of Haemophilus influenzae type b meningitis after routine childhood immunisation with conjugate vaccines. Lancet 1992; 340: 592–94. 3 Adams WG, Deaver KA, Cochi SL, et al. Decline of childhood Haemophilus influenzae type b (Hib) disease in the Hib vaccine era. JAMA 1993; 269: 221–26. 4 Lagos R, Levine OS, Avendano A, Horwitz I, Levine MM. The introduction of routine Haemophilus influenzae type b conjugate vaccine in Chile: a framework for evaluating new vaccines in newly industrialized countries. Pediatr Infect Dis J 1998; 17 (suppl 9): S139–48. 5 Adegbola RA, Usen SO, Weber M, et al. Haemophilus influenzae type b meningitis in The Gambia after introduction of a conjugate vaccine. Lancet 1999; 354: 1091–92. 6 Wenger JD, DiFabio J-L, Landaverde JM, Levine OS, Gaafar T. Introduction of Hib conjugate vaccines in the non-industrialised world: experience in four “newly adopting” countries. Vaccine 2000; 18: 736–42. 7 Bijlmer HA. World-wide epidemiology of Haemophilus influenzae meningitis; industrialized versus non-industrialized countries. Vaccine 1991; 9 (suppl): S5–9. 8 Adegbola RA, Mulholland EK, Falade AG, et al. Haemophilus influenzae type b disease in the Western Region of the Gambia: background surveillance for a vaccine efficacy trial. Ann Trop Paediatr 1996; 16: 103–11.
150
10
11
12
14
15
16
17
18
19
20 21 22 23
24 25
Bijlmer HA, van Alphen L, Greenwood BM, et al. The epidemiology of Haemophilus influenzae meningitis in children under five years of age in The Gambia, West Africa. J Infect Dis 1990; 161: 1210–15. Goetghebuer T, West TE, Wermenbol V, et al. Outcome of meningitis caused by Streptococcus pneumoniae and Haemophilus influenzae type b in children in The Gambia. Trop Med Int Health 2000; 5: 207–13. Greenwood BM. Epidemiology of acute lower respiratory tract infections, especially those due to Haemophilus influenzae type b, in The Gambia, West Africa. J Infect Dis 1992; 165: S26–28. Mulholland K, Hilton S, Adegbola R, et al. Randomised trial of Haemophilus influenzae type-b tetanus protein conjugate vaccine for prevention of pneumonia and meningitis in Gambian infants. Lancet 1997; 349: 1191–97. Adegbola RA, Mulholland EK, Secka O, Jaffar S, Greenwood BM. Vaccination with a Haemophilus influenzae type b conjugate vaccine reduces oropharyngeal carriage of H influenzae type b among Gambian children. J Infect Dis 1998; 177: 1758–61. Milligan PJM, Njie A, Bennett S. A comparison of two cluster sampling methods for health surveys in developing countries. Int J Epidemiol 2004; 33: 469–76. Cutts FT, Zaman SMA, Enwere G, et al. Efficacy of nine-valent pneumococcal conjugate vaccine against pneumonia and invasive pneumococcal disease in The Gambia: randomised, double-blind, placebo-controlled trial. Lancet 2005; 365: 1139–46. Decker MD, Edwards KM, Bradley R, Palmer P. Comparative trial in infants of four conjugate Haemophilus inlfuenzae type b vaccines. J Pediatr 1992; 120: 184–89. Bulkow LR, Wainwright RB, Letson GW, Chang SJ, Ward JI. Comparative immunogenicity of four Heamophilus influenzae type b conjugate vaccines in Alaska Native infants. Pediatr Infect Dis J 1993; 12: 484–92. Waggoner-Fountain L, Hendley O, Cody EJ, Perriello VA, Donowitz LG. The emergence of Haemophilus influenzae types e and f as significant pathogens. Clin Infect Dis 1995; 21: 1322–24. Ward JI. Editorial response: invasive infections due to Haemophilus influenzae serotype F (Hif) Is Hif an emerging pathogen? Clin Infect Dis 1996; 22: 1077–78. Baer M, Vuento R, Vesikari T. Increase in bacteremic pneumococcal infections in children. Lancet 1995; 345: 661. Gieseche J, Fredlund H. Increase in pneumococcal bacteremia in Sweden. Lancet 1997; 349: 670–99. Jordens JZ, Slack MPE. Haemophilus influenzae: then and now. Eur J Clin Microbiol Infect Dis 1995; 14: 935–48. Dagan R, Fraser D, Roitman M, et al. Effectiveness of a nationwide infant immunization program against Haemophilus influenzae type b. Vaccine 1999; 17: 134–41. Garner D, Weston V. Effectiveness of vaccination for Haemophilus influenzae type b. Lancet 2003; 361: 395–96. 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.
www.thelancet.com Vol 366 July 9, 2005