In developing countries, Streptococcus pneumoniae is a leading cause of pneumonia in hospitalized infants and young children, and is an important agent in bacterial meningitis and other invasive infections.1 In clinical trials and studies in the United States,2–4 South Africa5 and The Gambia,6 pneumococcal conjugate vaccines (PCV) have shown good protection against invasive pneumococcal disease (IPD) caused by the vaccine serotypes in infants, as well as against radiologically defined pneumonia.5–7 In The Gambia, PCV reduced all-cause hospital admissions and all-cause mortality.6 PCV is regarded as a potential tool to reduce the global burden of pneumonia and infant and childhood deaths.8
A recent meta-analysis has summarized evidence on most of the relevant controlled efficacy trials on PCV.9 Data are lacking, however, on PCV efficacy in Asia where 3/4 of the world’s infant deaths from pneumonia are estimated to occur.10 Pneumonia remains a leading cause of death among infants and young children in the Philippines.11 Pneumococcus is thought to be the commonest bacterial agent implicated in the etiology of pneumonia but isolation rates from the blood are low.12,13 We report an individual-randomized, double-blind, placebo-controlled efficacy trial of an investigational, nonadjuvanted 11-valent PCV (11PCV)14 against radiologically proven community-acquired pneumonia in infants and children less than 2 years of age in Bohol, the Philippines (ISRCTN 62323832).
SUBJECTS AND METHODS
The study was conducted in 6 municipalities (Tagbilaran City, Dauis, Panglao, Baclayon, Cortes, and Balilihan) in Bohol Province in central Philippines. This is a predominantly rural agricultural area covering 357 kms2 with a population of 149,000 at the start of the study in 2000. There are 3 recognized seasons: the hot season from March to May, the rainy season from June to November, and the cold season from December to February. There is no malaria and practically no HIV in Bohol. In 2005, the Gross National Income of the Philippines was 1300 USD per capita. The infant mortality rate (IMR) in Bohol during years 1999 to 2002 was 28/1000 live births the major causes of deaths being pneumonia and diarrhea.15 The study followed Good Clinical Practice guidelines from the International Center for Harmonization and was approved by the ethics review boards of the Research Institute for Tropical Medicine, the Philippines (RITM), and the National Public Health Institute of Finland (Kansanterveyslaitos; KTL). An independent data and safety monitoring board monitored the trial.
Study Design and Overall Implementation
The trial was a randomized, placebo-controlled, double-blind trial (ISRCTN 62323832). It required close collaboration with government and private health services in Bohol in the recruitment, enrolment, and vaccination of infants through municipal primary health services, and in surveillance for primary and secondary end points, which was conducted in the Bohol Regional Hospital (BRH) and in private hospitals. The local government had operationalized the ARI Case Management program, although antibiotics were often missing from the primary health care centers. Children often contacted the health services more than once in the course of a disease episode. We distinguished between disease events, which represented a single contact, and disease episodes, which could represent multiple contacts. Because the incidence of invasive disease was low, the study lacked a highly specific end point and episodes were categorized in several different ways.16
Identification, Recruitment, and Enrolment of Trial Participants
The study nurses started the process of recruitment by giving information about the study to pregnant mothers who came in for prenatal check-up at health centers and to mothers who had just given birth at hospitals. They then followed them up at their homes for detailed recruitment. The study nurses further searched the civil registry for information about births that may have not been included in the recruitment list. Enrollment and vaccination of study participants took place in 48 government primary health care centers when mothers or caretakers brought their infants for first vaccination. If the mother agreed to participate in the trial, an informed consent was signed, and the child was enrolled if (a) the family was expected to remain in the study area for the following 2 years or till the end of December 2004; (b) the infant was healthy; and (c) the infant did not meet exclusion criteria (receipt of a first dose of diphtheria- tetanus-whole cell pertussis vaccine (DTwP); a rectal temperature ≥38.5°C; neurologic disease; history of hospitalization for and/or treatment for immune suppression; or enrolment in another clinical trial). In a nested study at 3 health centers, the study nurses monitored subsets of the enrolled children for reactogenicity and immunogenicity of the study vaccine.
Vaccines and Allocation to Intervention
The investigational 11PCV prepared and manufactured by sanofi pasteur contained 1 μg of S. pneumoniae capsular polysaccharide conjugated to tetanus toxoid for types 1, 4, 5, 7F, 9V, 19F, and 23F; and 3 μg of polysaccharide of types 3, 14, and 18C conjugated to diphtheria toxoid; and 10 μg of polysaccharide of type 6B conjugated to diphtheria toxoid. Saline served as placebo. Both the 11PCV and placebo were contained in prefilled, ready-to-use, glass syringes, and looked identical. Study vaccines were given concomitantly with sanofi pasteur vaccines (Lyon, France): DTwP mixed with Haemophilus influenzae type b (PRP-T) conjugate vaccine, oral polio vaccine, and Hepatitis B (Heb B, MedTest, Korea). Measles vaccine was given at 9 months of age, whereas Bacillus Calmette-Guerin vaccine was given within a week after birth or during the first DTwP vaccination. Three doses of the study vaccines and DTwP/PRP-T, HepB, and OPV vaccines were given following the Philippine Expanded Programme on Immunization schedule (ie, officially 6, 10, and 14 weeks of age) starting at 6 weeks to <6 months of age.
Block randomization was used to allocate 11PCV and placebo vaccines at the individual level. A list containing random permutations of the letters A to F was generated by sanofi pasteur using SAS software (SAS Institute, Inc, Cary, NC); 3 of the letters were allocated to the 11PCV vaccine and 3 to the placebo. The correspondence between letters and vaccine type was unknown to researchers or subjects. After a child was enrolled, he or she was allocated to the next available letter on the list. Vaccine letter codes were concealed from the study nurse until uncovered at the time of allocation. The corresponding lettered solution was administered to the infant and recorded on the case report form. The code list was kept in sealed envelopes at 3 sites only: sanofi pasteur, the safety monitor of DSMB, and the Independent Technical Services Unit located at the National Health and Medical Research Council Clinical Trials Centre, based at the University of Sydney (Australia).
Either 11PCV or placebo was injected intramuscularly into the anterolateral aspect of the right thigh in a double masked manner; concomitantly, routine vaccines were injected into the left thigh. The child’s randomization code was recorded on a sticker affixed to the child’s vaccination card, which was removed when the child’s primary series of 3 doses was complete. The vaccination card with the child’s name and the study ID number were used to identify study subjects.
Definition of Study End points and Serious Adverse Events
Community-acquired pneumonia (CAP) was defined as pneumonia with onset either in the community or in hospital but less than 72 hours after admission into hospital. Cases of nosocomial pneumonia (ie, pneumonia identified after child had been admitted for 72 or more hours) were excluded from all analysis.
The primary trial end point was radiologically-defined CAP using the WHO vaccine trialists’ standardized guidelines (ie, WHO defined primary end point pneumonia, WHO-PEP)17; it was defined as the presence of a dense opacity that could be a fluffy consolidation of a portion of a lobe, a whole lobe, or the entire lung, often containing air bronchograms and sometimes associated with pleural effusion in the lateral pleural space associated with a pulmonary infiltrate or an effusion large enough to obscure such an opacity.
Secondary end points were clinical CAP either hospitalized or nonhospitalized, culture proven invasive disease with a vaccine type-specific pneumococcus, and serious adverse events (SAEs). Clinical pneumonia, which was classified by WHO severity grade18 was present if a child had a history of cough and/or difficult breathing of less than 2 weeks duration, and presented with (a) increased respiratory rate (rate ≥60/min if age <2 months, ≥50/min if age 2–11 months and ≥40/min if age 12–59 months); (b) lower chest wall indrawing (severe pneumonia); or (c) cyanosis and/or inability to feed or drink (very severe pneumonia). For the purposes of this study, IPD was defined either as bacteraemia or culture proven meningitis.
An SAE was defined as any untoward medical occurrence after immunization that resulted in death, was life threatening, required in-patient hospitalization or prolonged existing hospitalization, or resulted in persistent or significant disability/incapacity. The WHO Adverse Reaction Terminology (WHO-ART) was used to code SAE diagnoses.19 In WHO-ART, the definition of pneumonia is very broad containing lower respiratory tract conditions such as nonspecified viral and bacterial pneumonia, and acute bronchiolitis.
Surveillance for Clinical Episodes
There were 8 hospitals in the study area. The BRH is a 250-bed tertiary government facility, and main clinical referral center on the island. Of the 7 private hospitals, 3 were equipped to provide good quality chest radiographs, and willing to participate in clinical surveillance.17,20 To be able to capture as many primary end points as possible, all mothers or caregivers were encouraged to bring their children to the health center or BRH OPD/hospital for any illness that they experienced at any time during the study. At the BRH and 3 private hospitals in Tagbilaran City, patients less than 2 years of age were screened for CAP and clinical suspicion of sepsis and meningitis. Episodes were considered as “new episodes” if they began at least 14 days after cessation of the earlier one.
Children admitted at the BRH with signs and symptoms of clinical pneumonia, sepsis, or meningitis were examined and investigated; the protocol included a blood culture and chest x-ray for all episodes, and CSF culture when indicated. Children attending the BRH outpatient department (BRH-OPD) who were not admitted to the wards and children admitted to the 3 private hospitals were investigated for signs of clinical pneumonia only, including chest x-rays, which were taken if the clinician suspected pneumonia. At the BRH, antibiotic treatment, blood counts, bacterial cultures, and chest x-rays were free of charge for enrolled children if they had clinical pneumonia, sepsis, or meningitis. At the 3 private hospitals where only clinical pneumonia cases were screened for possible vaccine efficacy (VE) end points, only chest x-rays were free. No transport costs were given to the caregivers.
Study personnel were permanently assigned to the hospitals involved with case ascertainment. They were trained in the WHO pneumonia-management algorithm18 and were monitored weekly for accuracy of measurement of respiratory rate, chest indrawing, and other signs of pneumonia. Interobserver variation in pneumonia severity assessment was kept to less than 10% because of the intensive monitoring during the study period. In the BRH admitting section, study nurses and physicians were available 24 hours a day all year round. They took clinical histories, examined patients, and ordered laboratory and radiologic investigations according to study protocols. The 3 private hospitals were visited by the study nurses twice a day from Monday to Friday and once a day on Saturdays to collect clinical data from study patients. Trial personnel were not involved in therapeutic decision-making.
Surveillance for Serious Adverse Events
The BRH, the 3 private hospitals participating in the clinical event surveillance and 4 other private hospitals were kept under surveillance for SAEs. Information about hospitalizations occurring outside the trial area was obtained from guardians. Study nurses scrutinized church records and civil registers for death events weekly, and visited the homes of children who had left hospital against medical advice to determine the outcome of the episode. On vaccination days, study nurses questioned health staff about SAEs that had not resulted in hospitalizations and about any deaths in the area. A safety monitoring team of 4 local physicians and 1 of a pool of 5 Finnish physicians reviewed SAE data every week to allocate WHO-ART codes and to determine a possible causal relationship between events and vaccination. The study nurses visited homes to determine the status of enrolled children on their second birthday or on December 31, 2004, whichever was the sooner.
Processing, Reading, and Arbitration of Radiographs for the Primary End Point Detection
A quality assurance system was set up for reading chest radiographs.20 Three clinicians at RITM and 1 Australian pediatric radiologist were trained in the use of WHO standardized radiograph guidelines and software every time a batch of radiographs was to be read. Two study readings were made (1 by the pediatric radiologist and the other by 2 of 3 RITM clinicians reading the films together and arriving at a consensus opinion). We scanned all chest films using a Sierra film digitizer (Vidar Systems Corporation, Herndon, Virginia, 2000). Scanned images of all discordant and 10% of concordant films for the WHO-PEP and an equal number of films read as having no WHO-PEP were sent to WHO for quality control. The final reading was taken to be that of the readers when their assessment was concordant, or when discordant, the final reading was that of the WHO.
After establishing study specific bacteriological surveillance procedures,21 isolation of S. pneumoniae from blood and CSF was performed using standard culture methods. Serotyping of pneumococcal isolates was done with the Quellung checker board method22 at RITM with quality assessment at KTL. A standardized enzyme immunoassay was used at KTL to determine the antibody response to each of the 11 pneumococcal polysaccharides in serum samples in the nested study.23
Statistical Analysis and Sample Size
The primary end point for the trial was the occurrence of the first episode of community acquired radiologically defined pneumonia within a child. Subsequent episodes within the same child were not counted in the primary analysis. The sample size was calculated in early 2000 to provide a 95% confidence interval that excluded 15% efficacy, with 80% power. At the estimated pretrial incidence rates of radiologically defined pneumonia (ie, 2383 of 100,000), it was estimated that 236 cases would be required, requiring the enrolment of 12,190 children. It was estimated that this would be achieved within 48 months, ie, including 12 months of follow-up after stopping enrolment after 36 months.
The primary analysis was performed per protocol (PP): only children who received all 3 doses of the study vaccine (letter codes as randomized) with a minimum interval of 21 days between doses were included, and their observation time for end points began 14 or more days after the third dose and finished at the date of exit from the trial. The intent-to-treat analysis (ITT) included children who had received at least 1 dose of the study vaccine. Their observation time started immediately after the first vaccine dose of a child and lasted up to their date of exit. A child exited from the study at 24 months of age, or at withdrawal, or at death, or when the clinical phase of the trial came to its end in December 31, 2004.
Rates of experiencing an end point episode were calculated per 100,000 person-years of observation, within each vaccine group. Efficacy of the vaccine in preventing a child experiencing an episode was calculated as 100 (1-RR)%, where RR is the rate ratio 11PCV: placebo. A P value for the difference between rates and 95% confidence intervals (95% CI) for the RR were calculated using a Poisson regression model. The 95% CI for efficacy was obtained by applying the transformation 100 (1-RR) to the 95% CI limits for RR.
Supplementary analyses, using episode counts for each child, and time to first event were performed using negative binomial regression and Cox proportional hazards models, respectively. No adjustment for multiple testing was performed when analyzing the SAE findings.
Data entry used an ACCESS/SQL database management system. SAS version 9.1 (SAS Institute, Inc) was used for data analyses.
Role of Funding Sources
Funding of the study was provided by all members of the ARIVAC consortium (ie, RITM, KTL, University of Queensland, University of Colorado, sanofi pasteur), and the European Commission DG Research INCO program, the Academy of Finland, the Finnish Ministry of Foreign Affairs, the Finnish Physicians for Social Responsibility, GAVI’s PneumoADIP, PATH and WHO. Of these, only the consortium members were active in all phases of the study: the design; the collection, analysis, and interpretation of data; the writing of an article, and decision to submit it for publication. PATH reviewed a final draft before submission. The WHO DSMB reviewed safety data and interacted with the consortium throughout the trial itself. The Australian NHMRC supported preparation of the study area.
Recruitment commenced in July 2000. Enrolment of the planned number of the study participants was completed December 18, 2003; follow-up of children ended December 31, 2004. Of 15,593 infants born in the study area, 2406 were not eligible for the trial (Fig. 1, Supplemental Digital Content 1, http://links.lww.com/A956). Of 12,194 enrolled children, 12,191 received the allocated intervention. Overall, 98.7% of subjects received 3 doses of study vaccine. The median age at vaccination was 1.8, 2.9, and 3.9 months for first, second, and third doses, respectively. Vaccine and placebo groups did not differ in key baseline indicators (Table 1, Supplemental Digital Content 2, http://links.lww.com/A957). Vaccinees were evenly distributed to the 2 groups over municipality, birthplace, birth attendant, number of live siblings, maternal age at enrolment, mother’s education, maternal tetanus toxoid status, completion rates (data not shown), and person-years of follow-up.
In the course of the trial, 3283 clinical events of interest were recorded. After application of rules and definitions to clinical events of interest, 3074 clinical episodes of interest were analyzed for VE. Chest radiographs were taken in 2789/3074 (91%) of episodes; 2742 (98%) of these were readable. A description of these episodes among the hospital admitted and OPD attended children stratified according to the clinical severity of the infection and chest radiographic finding is Table 2.
VE Against Primary and Secondary End Points
In per protocol analysis (Table 3), a first episode of WHO-PEP was identified in 93 children (1040 cases/100,000 PY) in the 11PCV group and 120 (1349 cases/100,000 PY) in the placebo group with VE of 22.9% (95% CI: −1.1 to 41.2; P = 0.06). VE was 34.0% (95% CI: 4.8 to 54.3; P = 0.02) in children <1-year-old; VE was 2.7% (95% CI: −43.5 to 34.0; P = 0.88) in children 12 to 23 months. The P value for interaction between age groups (ie, difference in efficacy) in the PP analysis was 0.14.
In the ITT analysis, the number of children with end point episodes was approximately 20% higher: VE was 16% (95% CI: −7.3 to 34.2; P = 0.16) overall; VE was 19.8% (95% CI: −8.8 to 40.8; P = 0.15) in children <1-year-old and 4.4% (95% CI −40.5 to 35.0; P = 0.81) in children 12 to 23 months (Table 4). The difference in VE between PP (Table 3) and ITT (Table 4) analyses is due to a nonsignificant difference in the number of WHO-PEP episodes in the vaccine group in the period prior to the start of the PP time window. In total, the vaccine group experienced 32 episodes of WHO-PEP, compared with 23 in the placebo group in the pre-PP period (P = 0.23). The excess of events was largely accounted for by episodes occurring between doses 2 and 3 (16 vs. 8, P = 0.10).
In per protocol analysis of secondary end points (Table 3), VE was 0.1% against all clinical pneumonia. VE was 3.7% against hospitalized pneumonia and −3.8% against nonhospitalized pneumonia. None of these was statistically significant. VE was −3.0% against nonsevere pneumonia, 8.4% against severe pneumonia, and −27.0% against very severe pneumonia. These differences were nonsignificant. The ITT analysis also showed low, nonsignificant VE (Table 4).
Results of the Cox proportional hazards model were almost identical to those given above; the negative binomial regression produced slightly lower estimates of efficacy, for both ITT and PP analyses.
The 1213 blood cultures (625 from 11PCV, and 588 from placebo recipients) and 77 CSF cultures (35 from 11PCV, and 42 from placebo recipients) from trial participants admitted to BRH yielded 6 episodes of IPD (29 cases/100,000 PY; 2 meningitis, 4 sepsis cases); 5 of these were in the per protocol data set (pneumococcal serotypes: 5, 6B, 14, 18A, and 38); 3 of these were due to vaccine serotypes (2 in 11PCV group, serotypes 5 and 6B, and 1 in the placebo group, serotype 14), and 1 due to vaccine related serotype (18A in the placebo group). The only nonvaccine serotype 38 was in the placebo group. In the ITT data set, there was an additional serotype 5 in the 11PCV group.
In all, 1111 children were enrolled into the nested immunogenicity study. One month after the third dose of study vaccine, at 18 weeks of age, the fully vaccinated 11PCV recipients had significantly (P < 0.0001) higher geometric mean antibody concentrations than the placebo recipients against each of the 11 vaccine serotypes (Fig. 2). By the time measles vaccination was given at median 9.5 months of age in the nested study children, the antibody concentrations had decreased but were still significantly higher in 11PCV than in placebo recipients (P < 0.0001).
At 18 weeks, 93% to 100% of 11PCV recipients achieved a threshold concentration of 0.35 μg/mL for 9 serotypes, 76% for 6B, and 87% for 23F. At 9 months of age, the proportions of the 11 serotypes varied between 24% and 97%.
Reactogenicity and Safety of 11PCV
A total of 366 children were monitored in the nested study for reactogenicity by use of a diary card. We observed no significant differences in the occurrence of local or systemic reactions after the first and second dose of study vaccine between 11PCV and placebo groups. After the third dose there were significantly more cases of any transient injection site erythema in the 11PCV than the placebo group; 35 (18.6%) versus 12 (6.72%), respectively (P = 0.0007).24
The SAE surveillance carried out during the whole trial period resulted in the identification of 4553 SAEs: 2324 (51%) in 11PCV and 2229 (49%) in the placebo group. Out of the 6097 11PCV recipients, 1624 (27%) had at least 1 SAE. Out of the 6094 placebo recipients, 1596 (26%) had at least 1 SAE. SAEs of 43.5% were detected in the BRH, 55.8% in the 7 private hospitals, and <1% in health centers; 33.8% of SAEs were categorized as pneumonia (WHO-ART Code 528) and 29.3% as gastroenteritis (WHO-ART Code 293). Four thousand two hundred twenty-nine of the SAEs (92.9%) were hospital admissions. Differences between 11PCV and placebo groups were nonsignificant in the study as a whole when time period from vaccination or different diagnostic categories were not taken into account.
However, SAEs from all causes were significantly increased in the 11PCV group compared with the placebo group within 7 days (RR = 2.39; P = 0.02), and 28 days of the first dose of 11PCV (RR = 1.58; P = 0.005); this phenomenon was not observed after the second or third dose of vaccine. SAEs from pneumonia (WHO-ART 528) were increased in the 11PCV group in comparison to the placebo group within 7 days (RR = 3.66; P = 0.05) and within 28 days of the first dose (RR = 1.43; P = 0.09), and also within 28 days of the second dose (RR = 1.62, P = 0.04). SAEs from causes other than pneumonia were also increased in the 11PCV group in comparison to the placebo group within 28 days after the first dose (RR = 1.85; P = 0.02). This was a heterogeneous group of diagnoses and included acute gastroenteritis, and febrile and nonfebrile seizures but the effect could not be attributed to any particular diagnostic category, probably because the number of cases was too small. Further details on the SAEs, especially in relation to viral involvement, will be published in future articles.
Sixty-four children died during the trial: 30 were in the 11PCV group and 34 in the placebo group. Out of the 64 deaths, 42 (66%) died in the hospital, 19 were home deaths, and 3 died on the way to the hospital. Sixteen children died of pneumonia: 7 were in the 11PCV group and 9 were in the placebo group. None of these differences was statistically significant. On the basis of surveillance for all deaths in the 6 municipalities, we estimate that, during the time of the trial, IMR in the study area as a whole was 25.3/1000 and that IMR in the trial population itself was 16.4/1000. (As IMR includes children from birth to 6 weeks of age, the estimate for the trial population includes infants who would have been enrolled had they survived to 6 weeks of age).
We have shown in this double-blind, individual-randomized, placebo-controlled trial in the Philippines that an investigational 11-valent diphtheria-tetanus toxoid conjugated PCV given in 3 doses prevented 22.9% of radiologically defined pneumonia (WHO-PEP) among children under 2 years of age.
The Philippines setting differs from the 2 earlier trials in Africa5,6 in important ways: the trial was conducted in a low-income in-transition country in Asia, where IMR was low (28/1000 live births),15 where access to health care and antimicrobials was good, and the population unaffected by HIV or malaria. Despite these differences, the VE of 22.9% (P = 0.06) in the per protocol analysis, while not meeting the conventional P < 0.05 for statistical significance, was of the same order of magnitude as in studies in South Africa among HIV-negative children (VE = 25%, 95% CI: 4–41). The similarity of the VE obtained in these very different study populations suggests that a similar proportion of all pneumonias in these settings identified as WHO-PEP is of pneumococcal etiology.17
An age-stratified analysis of the primary end point was included in our original plan of analysis to allow comparisons with other trials. We found an unexpectedly strong age dependence of the VE; an effect in children <12 months old (VE = 34.0%) and the comparative lack of an effect in those 12 to 23 months of age (VE = 2.7%). The difference in VE between the age groups was not statistically significant (P = 0.14); the study was not, however, powered to detect an age-VE interaction. A corresponding age-specific difference in VE was not detected in the Gambian trial,6 but was detected in North California.25 The analysis of the age-specific estimates for our study was planned a priori and the finding in young children achieved statistical significance.
The 11PCV used in this trial was an investigational vaccine administered concomitantly with DTwP that included the same serotypes as the 9-valent PCV, with the addition of types 3 and 7F. It was well-matched to the distribution of serotypes causing IPD in the Philippines.13,26 It was shown to induce a significant antibody response to each of the 11 serotypes with similar magnitudes of serotype-specific antibody responses and proportions reaching putative thresholds of 0,35 μg/mL seen with the 7- and 9-valent vaccines products.2,27–30 The persistence of antibodies up to 9 months of age was similar to that reported for the 9PCV in South African infants.29 In contrast, an 11PCV with similar concentration of conjugated pneumococcal antigens administered concomitantly with acellular pertussis vaccine, has been shown to have less immunogenicity than when administered with DTwP in a separate study.31 Our study provides needed evidence that tetanus and diphtheria toxoid based PCVs are immunogenic and efficacious when given concomitantly with wP vaccine, information important for new vaccine producers especially aimed at low income country markets.32
We observed a small but significant increase in all SAEs after the first dose of 11PCV due partly to an increase in pneumonia (WHO-ART 528) but also to a mix of conditions in other diagnostic categories; SAEs from pneumonia were increased also in the first 28 days after the second dose of 11PCV. The WHO-ART category of pneumonia overlaps with, but is broader than, the trial categories of PEP and clinical pneumonia. The South African trial of PCV reported a transient increase in the incidence of respiratory syncytial virus pneumonia 1 to 8 days after vaccination; the authors also reported an increase in reactive airway disease not related to the time of vaccination.5 The Gambian trial reported more total outpatient consultations among the vaccine recipients for most clinical categories after the first dose of 9PCV, but no differences were observed after subsequent doses.6 Both the SAE analysis and the clinical end point analyses showed an early excess of events when comparing the 11PCV group to the control group in the ITT compared with the PP population. As shown above, much of this excess was due to events occurring before the full development of immunity (ie, 14 days after the third vaccine dose set, as customary, as the starting point of PP observation). The data are also consistent with an increased incidence of early viral respiratory disease and/or viral bacterial interactions in the vaccinees. These aspects are the subject of an ongoing study, and will be published separately.
Given that pneumonia is the major cause of death in under fives in developing countries, an important public health question is the effect of 11PCV on pneumonia disease burden: would PCV also prevent an important part of all clinical pneumonia? We found that only a small proportion of clinically diagnosed pneumonia in this setting (13% in the placebo group) was radiologically defined WHO-PEP. Three previous trials2,5,6 have resulted in point estimates of 6% to 7% VE against clinical pneumonia. However, only in the study in The Gambia has this been statistically significant, and then only seen in those children who also had WHO-PEP.6 In the Philippines, we observed no VE against clinical pneumonia but the point estimate fell within the CI bounds of the 3 other trials. Although 11PCV had no effect against clinical pneumonia, a formal meta-analysis done on this end point (9, being updated) showed that the pooled estimate from the results of these 3 trials and this study (6% [95% CI, 2–9]) was highly significant (P = 0.0006); no statistical heterogeneity was observed. In an exploratory analysis, we stratified the Philippines data using both the WHO definitions and out-patient visits versus hospitalizations for pneumonia but could not demonstrate any consistent relationship between VE and clinical severity.
Invasive disease caused by vaccine types of pneumococci (IPD) would be the most specific end point for vaccine studies. Unfortunately its sensitivity is low, resulting in underestimation of the real burden of potentially preventable pneumococcal disease. Accordingly, the primary end point in the present study was radiologically defined pneumonia, identified by the best available criteria to be as specific as possible for pneumococcal pneumonia.17 IPD was included in the study protocol as one of the secondary end points, knowing that it would be unlikely to give a statistically significant VE because of the low rate of IPD in children in the area.12,13 We paid, however, special attention to the sampling and laboratory procedures, but the yield of positive blood cultures was very low, making any inferences to vaccine effect impossible. The overall rate of 29/100,000 child-years of IPD in the study population is indeed low, compared with rates of 100 to 380/100,000 reported in Africa and USA5,6,33 although low rates have been reported for example in Hong-Kong (18/100,000) and Finland (45/100,000).34,35 In each case, the reasons for the low rates have remained a matter of speculation only: insensitive indications for ordering blood or CSF cultures, defective laboratory techniques, and administration of antimicrobials before the diagnostic samples were collected36 on one hand that all could make the low yields artificial, or on the other hand, truly decreased prevalence of severe pneumococcal disease because effective treatment of pneumococcal pneumonia had prevented its progression to IPD. Because of so many, and partially contradictory, factors affecting the results of blood culture studies their use for comparing populations and especially for drawing inferences regarding burden of disease should be taken with caution.
In summary, the tetanus-diphtheria toxoid conjugated 11PCV has shown similar efficacy against radiologically defined pneumonia (WHO-PEP) as the available 7PCV in very different geographic and epidemiological settings. This is good news for the emerging pharmaceutical industry developing new PCVs, some of which are tetanus toxoid based, and intended for the rapidly growing markets of the low and middle income countries. Our results give assurance to the use of WHO-PEP as feasible end point in clinical trials and fairly good proxy for measuring pneumococcal pneumonia—depending on the circumstances, a minimum one-fifth of WHO-PEP is caused and thus preventable by the presently chosen vaccine type pneumococci. The adverse event findings of the respiratory tract, contributing to the major differences between the ITT and PP analyses, is of concern and need further work before their significance can be understood. Finally, these results raise the question on the place of PCV against childhood pneumonia in countries with limited resources, IMR of under 50/1000, and low rates of IPD. This is especially important to discuss in the non-GAVI eligible countries, which the Philippines belongs to, where the Integrated Management of Childhood Illnesses is mostly functioning, and the cost of treatment of pneumonia at first level facilities is very much lower than the present cost of the two available PCVs. Decisions to include new vaccines into the national program are made based on the available estimates on disease burden, costs caused and averted by vaccination, country-specific health care priorities in relation to available resources, and political will. A large scale postintroduction impact study coupled with a sensitive and specific enough surveillance system designed to capture both invasive and noninvasive disease episodes in both children and adults in a representative Asian country would be welcome to address these key questions our findings raise.
This study is part of the research of the ARIVAC Consortium. We are indebted to the Consortium study team and the following collaborators–The Data Safety Monitoring Board: Kim Mulholland (chair); Keith Klugman; Mary Ann Lansang (local safety monitor); David Sack; Pratap Singhashivanon; Peter Smith; and Chongsuphajaisiddhi Tan; National Institute of Health and Welfare (THL; formerly National Institute of Public Health KTL): Tarja Kaijalainen, Kaisa Jousimies; Research Institute for Tropical Medicine (RITM): Vernoni Ermata Dulalia, Leilani T. Nillos, sanofi pasteur: S. Arnoux, F. Bailleux, S. B′Chir, E. Boutry, J. M. Chapsal, Y. Couedel, V. Delore, H. DyTioco, E. Feroldi, J. Lang, J. R. Maleckar, M. Moreau, R. Ryall, D. Schulz, D. Teuwen, S. Vital, and C. Zocchetti.
The ARIVAC Consortium thanks and acknowledges the participation of the infants, parents, staff, Local Government of the Province of Bohol and Local Government Units (LGUs) of Baclayon, Balilihan, Cortes, Dauis, Panglao and Tagbilaran City; staff of the Pathology and Pediatric Departments of the BRH, and the private hospitals Tagbilaran Community Hospital, Borja Family Clinic, Medical Mission Group of Hospitals, Ramiro Community Hospital, St. Jude Hospital, Englewood Hospital, and Tagbilaran Puericulture Center.
1. Levine OS, O’Brien KL, Knoll M, et al. Pneumococcal vaccination in developing countries. Lancet
2. Black S, Shinefield H, Fireman B, et al. Efficacy, safety and immunogenicity of heptavalent pneumococcal conjugate vaccine in children. Pediatr Infect Dis J
3. O’Brien K, Moulton L, Reid R, et al. Efficacy and safety of seven-valent conjugate pneumococcal vaccine in American Indian children: group randomised trial. Lancet
4. Whitney CG, Pilishvili T, Farley MM, et al. Effectiveness of seven-valent pneumococcal conjugate vaccine against invasive pneumococcal disease: a matched case-control study. Lancet
5. Klugman KP, Madhi SA, Huebner RE, et al. A trial of a 9-valent pneumococcal conjugate vaccine in children with and those without HIV infection. N Engl J Med
6. Cutts FT, Zaman SM, 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. (erratum: Lancet
. 2005;366:328) Lancet
7. Hansen J, Black S, Shinefield H, et al. Effectiveness of heptavalent pneumococcal conjugate vaccine in children younger than 5 years of age for prevention of pneumonia: updated analysis using World Health Organization standardized interpretation of chest radiographs. Pediatr Infect Dis J
8. Anonymous. Pneumococcal conjugate vaccine for childhood immunization. WHO position paper. Wkly Epidemiol Rec
9. Lucero M, Dulalia V, Parreno R, et al. Pneumococcal conjugate vaccines for preventing vaccine-type invasive pneumococcal disease and pneumonia with consolidation on x-ray in children under two years of age. Cochrane Database Syst Rev
10. Williams BG, Gouws E, Boschi-Pinto C, et al. Estimates of world-wide distribution of child deaths from acute respiratory infections. Lancet Infect Dis
11. Philippine Health Statistics. Department of Health National Epidemiology Center. Manila, the Philippines; 2000.
12. Tupasi TE, Lucero MG, Magdangal DM, et al. Etiology of acute lower respiratory tract infection in children from Alabang, Metro Manila. Rev Infect Dis
13. Lupisan S, Herva E, Sombrero LT, et al. Invasive bacterial infections of children in a rural province in central Philippines. Am J Trop Med Hyg
14. Wuorimaa T, Dagan R, Eskola J, et al. Tolerability and immunogenicity of an eleven-valent pneumococcal conjugate vaccine in healthy toddlers. Pediatr Infect Dis J
15. National Demographic and Health Survey. National Statistics Office, Manila, Philippines; 2003.
16. Puumalainen T, Quiambao B, Abucejo-Ladesma E, et al; for the ARIVAC Research Consortium. Clinical case review: a method to improve identification of true clinical and radiographic pneumonia in children meeting the World Health Organization definition for pneumonia. BMC Infect Dis
17. World Health Organization Pneumonia Vaccine Trial Investigators′ Group. Standardization of interpretation of chest radiographs for the diagnosis of pneumonia in children. Geneva, Switzerland: WHO; 2001. WHO/V&B/01.35.
18. World Health Organization Programme for the Control of Acute Respiratory Infections: acute respiratory infections in children: case management in small hospitals in developing countries. Geneva, Switzerland: WHO; 1990. WHO/ARI/90.5.
20. Arcay JD, Ocampo AF, Solis RK, et al. Radiology quality assurance in a developing country setting: the 11-valent pneumococcal conjugate vaccine trial, Bohol, Philippines. Vaccine
21. Herva E, Sombrero L, Lupisan S, et al. Establishing a laboratory for surveillance of invasive bacterial infections in a tertiary care government hospital in a rural province in the Philippines. Am J Trop Med Hyg
22. Sorensen UB. Typing of pneumococci by using 12 pooled antisera. J Clin Microbiol
23. Käyhty H, Åhman H, Rönnberg PR, et al. Pneumococcal polysaccharide-meningococcal outer membrane protein complex conjugate vaccine is immunogenic in infants and children. J Infect Dis
24. Ugpo J, Lucero MG, Williams G, et al. Reactogenicity and tolerability of a non-adjuvanted 11-valent diphtheria-tetanus toxoid pneumococcal conjugate vaccine in Filipino children. Vaccine
25. Black S, Shinefield H, Ling S, et al. Effectiveness of heptavalent pneumococcal conjugate vaccine in children younger than five years of age for prevention of pneumonia. Pediatr Infect Dis J
26. Capeding MR, Sombrero LT, Lucero MG, et al. Serotype distribution and antimicrobial resistance of invasive Streptococcus pneumoniae
isolates in Filipino children. J Infect Dis
27. Obaro SK, Enwere GC, Deloria M, et al. Safety and immunogenicity of pneumococcal conjugate vaccine in combination with diphtheria, tetanus toxoid, pertussis and Haemophilus influenzae type b conjugate vaccine. Pediatr Infect Dis J
28. Huebner RE, Mbelle N, Forrest B, et al. Immunogenicity after one, two or three doses and impact on the antibody response to coadministered antigens of a nonavalent pneumococcal conjugate vaccine in infants of Soweto, South Africa. Pediatr Infect Dis J
29. Huebner RE, Mbelle N, Forrest B, et al. Long term antibody levels and booster responses in South African children immunized with nonavalent pneumococcal conjugate vaccine. Vaccine
30. Madhi SA, Kuwanda L, Cutland C, et al. Quantitative and qualitative antibody response to pneumococcal conjugate vaccine among African Human Immunodeficiency Virus-infected and uninfected children. Pediatr Infect Dis J
31. Dagan R, Goldblatt D, Maleckar JR, et al. Reduction of antibody response to an 11-valent pneumococcal vaccine coadministered with a vaccine containing acellular pertussis components. Infect Immun
32. Berndt ER, Glennerster R, Kremer MR, et al. Advance market commitments for vaccines against neglected diseases: estimating costs and effectiveness. Health Econ
33. Robinson K, Baughman W, Rothrock G, et al. Epidemiology of invasive Streptococcus pneumoniae
infections in the United States, 1995–1998: opportunities for prevention in the conjugate vaccine era. JAMA
34. Ho PL, Chiu SS, Cheung CH, et al. Invasive pneumococcal disease burden in Hong Kong children. Pediatr Infect Dis J
35. Eskola J, Takala AK, Kela E, et al. Epidemiology of invasive pneumococcal infections in children in Finland. JAMA
36. Sombrero L, Sunico ME, Quiambao B, et al. Reliability of parental history of antibiotic use for Filipino children admitted with acute lower respiratory tract infection. Am J Trop Med Hyg