Global Pneumococcal Disease and Policies for Control

M. Deloria Knoll1, B.A. Nonyane2, C. Garcia1, O.S. Levine3, K.L. O’Brien1, H.L. Johnson4 1International Vaccine Access Center, Johns Hopkins Bloomberg School of Public Health, Baltimore MD, USA; 2International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore MD, USA; 3Global Development Program, Bill & Melinda Gates Foundation (formerly of JHSPH), Seattle WA, USA; 4Monitoring & Evaluation Policy & Performance, GAVI Alliance (formerly of JHSPH), Geneva, Switzerland


METHODOLOGICAL ISSUES IN THE ANALYSIS OF 'BEFORE AND AFTER' VACCINE STUDIES
Background and Aims: Evaluation of vaccine impact using 'before and after' data is complicated by temporal changes in factors which may confound or bias results. We addressed these issues in a study of the impact of pneumococcal conjugate vaccine in The Gambia. Methods: We analysed incidence of primary end-points in pre-and post-intervention periods and performed a cohort analysis. For the pre-and post-analysis, we defined the start of the intervention period as the date when coverage of ≥2 doses across an age range approached its maximum. Individuals may contribute time in both 'before' and 'after' periods (accounted for by a random effect for subject) to reduce loss of information. We defined potential confounders (e.g. season, malaria) and adjusted for changes in the comparison periods. We identified potential biases, such as changes in case ascertainment or the prevalence of serotypes known to vary markedly over time, and included these variables in stratified and multivariable modelling. End-points unrelated to vaccination (e.g. Gram-negative bacteraemia) were used to indicate potential confounding or bias. We estimated variance using robust methods. Children eligible for vaccination were included in the cohort analysis which may be interpreted as vaccine effectiveness at the individual level. To facilitate control of confounding we used a time-to-event approach specifying vaccination status and age as time-varying covariates. Results are presented in plenary session 6. Conclusion: Analytic techniques are available to improve the evaluation of new vaccines using 'before and after' data while taking into account potential confounding and bias. Background and Aims: Many pneumococcal (Sp) meningitis cases in the African meningitis belt result from the invasive serotype 1, which data from other areas indicate should present with relatively low mortality. Methods: We evaluated all age acute bacterial meningitis mortality data from healthcare structure based surveillance systems in northern Togo and western Burkina Faso. All patients had received care based on standard algorithms. Etiologic determination was performed via culture, antigen detection, or polymerase chain reaction (PCR) while serotype/group was determined by PCR or Quellung. Results: Among 4838 patients, the case fatality ratio (CFR) was 34% (275/808) for Sp, and 9% (56/639), 11% (26/229), and 15% (17/111) for serogroups A, W135, and X meningococcus, respectively. Sp CFR varied only modestly by study year (23% [2005] to 46% [2010]) and age group (31% [5-14 years] to 43% [50+ years]) with no particular trend; 56% of Sp deaths occurred from age 5-49 years. Of 355 (43%) evaluated Sp cases, 51% were serotype 1 and 66% were invasive (serotype/groups 1,4,5,7,14,18). CFR was 42% for serotype 1 versus other serotypes and for invasive versus non-invasive serotypes. Serotype 1 and invasive serotypes were not associated with lower CFR when stratified by country, age group, or study year. Antibiotic resistance did not predict death. Conclusion: In the African meningitis belt Sp meningitis mortality is higher than other etiologies, affects all age groups, shows no evidence of decreasing, and is not lower for serotype 1 or the broader category of invasive serotypes.

ISPPD-0492 Global Pneumococcal Disease and Policies for Control
Introduction: 23-valent pneumococcal polysaccharide vaccine (PPV) following pneumococcal conjugate vaccines (PCV) should increase protection however concerns exist regarding hypo-responsiveness and B-cell memory impairment after PPV. We evaluated the persistence of antibody and B-cell memory following PPV at 9 months of age in 3-5 year old children in Papua New Guinea. Methods: 150 children who had received PPV at age 9 mo (after PCV7 at 0-1-2 or 1-2-3 months or no PCV7) and 130 age-matched unvaccinated controls received a 0.1 mL PPV challenge dose at age 3-5 years and blood was collected pre-and 1 month post-challenge. Serotype-specific IgG for serotypes in PCV7 (VT), 2, 5 & 7F was measured by ELISA. B-cell memory for serotypes 1, 2, 7F, 14, 19A, 19F & 23 was measured at age 10 and 18 months (prior-PPV groups) and pre-challenge by polyclonal B-cell stimulation and detection of antibody forming cells on polysaccharide-coated plates by ELISpot. Results: Post-PPV, memory B-cells numbers increased from age 10 months to 3-5 years for 8/9 serotypes (not 23F) and did not differ to PPV naïve controls. Pre-challenge, IgG levels were high, with no significant differences between groups in GMCs or proportions ≥1 μg/mL (range 58-95%). IgG increases after PPV-challenge were modest (1.2 to 2.8-fold) with no significant differences in GMCs, fold-rises or % ≥1 μg/mL between groups. Increased pre-challenge antibody concentrations were associated with a decreased IgG response to challenge. Conclusion: We found no evidence of hypo-responsiveness or impairment of memory B-cells in PNG children at 3-4 years of age after PPV at age 9 months.

OTITIS MEDIA AND ITS SEQUELAE IN KENYAN SCHOOL CHILDREN
E. Simões 1 , I.M. Macharia 2 , F. Kiio 2 , P. Carosone-Link 1 , S.N. Ndegwa 2 , J. Ayugi 2 1 Dept. of Global Health, University of Colorado at Denver, Denver, USA; 2 Dept. of Surgery, University of Nairobi Kenyatta National Hospital, Nairobi, Kenya Background and Aims: Chronic suppurative otitis media (CSOM), a sequel of acute otitis media (AOM) is a major cause of preventable hearing loss in children in developing countries. In Africa there are few recent studies on the prevalence of AOM and its sequelae. Obtaining representative Kenyan data on the point prevalence of AOM and its sequelae, otitis media with effusion (OME), CSOM and hearing impairment was the goal. Methods: Study subjects were children aged 2 to 15 years enrolled from randomly selected preprimary and primary schools. With parental/guardian consent, subjects had a questionnaire administered, otoscopy and tympanometry done, and audiometry performed on those with ear problems detected on these examinations. Results: Of the 13,109 children examined, 75% were from rural schools. The prevalence's of CSOM and OME were 15/1000 children and AOM was 7/1000. Apart from retraction of the tympanic membrane (urban 0.3%, rural 0.1%, p < 0.001), there was no statistically significant difference between the prevalence of other middle ear disorders in rural and urban school children. However in rural children, those from the Rift Valley had a significantly higher rate of CSOM 24/1000 than other regions 12/1000; p < 0.0001. The age of onset of ear discharge was before 3.5 years in 50% of children and before 6 years in 75%. Conclusion: Since a significant burden of AOM sequelae occur in Kenyan preschool and school children with the onset mostly, in the first 4 years of life, we question: would pneumococcal vaccines that prevent early recurrent AOM, prevent these nonreversible sequelae?
No conflict of interest pneumonia 2014 Volume 3 227 ISPPD-9 / pneumonia 2014 Mar 9-13;3:1-286 Background: The prevention of mother-to-child transmission (PMTCT) programme has been rapidly scaled-up in South Africa with a reported decrease in HIV transmission rates from 3.5% (2010) to 2.7% (2011). There are no published data quantifying specific increased risk or mortality of IPD in HIV-exposed uninfected (HEU) children.

Methods:
We conducted a cohort surveillance study in children <1 year of age with invasive pneumococcal disease (IPD), using data obtained from a national, laboratory-based, surveillance programme. HIV and outcome data were obtained at sentinel sites. Incidences in HIV-infected, HIV-uninfected, HEU and HIV-unexposed uninfected (HUU) children were calculated assuming similar annual HIV prevalence in tested and untested children. Results: We identified 1805 IPD cases in children <1 year of age from 2009 through 2012 from all sites. In children from sentinel sites (n=863), 89% (768/863) had known HIV status. The highest incidence of IPD was in HIV-infected (608/100,000 population) compared with HIV-uninfected children (31/100,000; crude relative risk 19.61 ). In addition, HEU children had higher incidence of IPD (61/100,000 population) than HUU children (22/100,000; crude relative risk 2.77 [95% CI 1.68-4.74]). Overall case-fatality ratio was high (27%, 234/863), differed between HIV-infected (32%, 75/231) and HIV-uninfected children (24%, 129/537, p=0.008); as well as between HEU (27%, 59/220) and HUU children (21%, 57/266, p=0.08). Discussion: The numbers of HIV-exposed children infected annually continue to drop due to improvements in PMTCT; however we have now shown that these exposed but uninfected children are at increased risk of IPD and mortality compared with HIV-unexposed children. Background and Aims: Invasive pneumococcal disease (IPD) is a serious public health problem in China, yet epidemiology data are limited, especially in adults. The study aimed to examine the serotype distribution of Streptococcus pneumoniae isolated from Chinese patients with IPD. Methods: The multicenter study collected S. pneumoniae positive IPD isolates from normally sterile sites in both pediatric (<18 years) and adult (≥18 years) patients in China between 2009 and 2013. Serotypes were analyzed by Chinese Center for Disease Control and Prevention using Capsule-quellung test. Results: A total of 110 isolates from 60 pediatric and 50 adult IPD cases were collected from 16 hospitals in 13 provinces in China. Overall, the most common serotypes in all ages were 19A, 19F, 23F, 3 and 14. The most frequent serotypes were 19A, 19F, 23F, 1, 5, 14, 6B and 6C in pediatric patients and 3, 19F, 19A, 23F, 14, 20 and 6A in adults. Interestingly, serotype 19A appeared significantly more frequent in infants (<2 years) than other age groups, whereas serotype 3 was the most common type in the elderly (>65 years). The common serotypes for main IPD diagnoses (invasive pneumonia, meningitis and sepsis) were 19A, 19F and 23F in invasive pneumonia and meningitis, and 19F, 19A, 5, 1 and 23F in sepsis. Conclusion: Serotypes 19A and 3 were the most prominent serotypes in pediatric and adult patients with IPD, respectively. Findings demonstrate the different serotype distribution among various age groups in Chinese population with IPD.

ASSESING THE ECONOMICS OF INVASIVE PNEUMOCOCCAL DISEASE IN OLDER CHILDREN AND ADULTS IN THE LATIN AMERICA AND CARIBBEAN REGION
D. Constenla 1 1 International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, USA Background: The economic burden of invasive pneumococcal disease (IPD) is substantial in the Latin American and Caribbean region. Objectives: To estimate the economic burden of IPD in the region from the healthcare system perspective and its impact on the gross domestic product (GDP) per capita. Methods: We contacted experts in five selected countries (Argentina, Brazil, Chile, Colombia and Uruguay) and asked them specific questions about the diagnosis and treatment of bacteremic pneumonia and meningitis in three age groups: 5-17 year olds, 18-64 year olds and 65+ year olds. Estimates of disease burden and costs were based on extensive literature review and public databases. Diseases considered included: pneumococcal pneumonia (inpatient/outpatient), and pneumococcal meningitis. Results: A total of 153 physicians responded to targeted questions. The total direct medical cost per treatment of pneumonia ranged from US$993 to US$3,132 in older children, US$1,274 to US$3,247 in adults and US$1,746 to U$3,535 in elderly, respectively. Overall, the costs of hospital stay accounted for 45% of the total treatment costs in these countries. Across the region, health care costs of IPD per country ranged from US$8.2 million to US$14.1 million, with higher costs incurred by the elderly due to higher level of resources used for treating the elderly. Healthcare spending for IPD in older children and adults as a percentage of GDP was estimated at 0.1%, compared to the reported 8-10% of GDP spent on healthcare overall. Conclusion: IPD poses a sizable economic burden among persons > 5 years in the LAC.

ECONOMICS OF INVASIVE PNEUMOCOCCAL DISEASE AMONG OLDER CHILDREN AND ADULTS IN LATIN AMERICA: RESULTS FROM A REVIEW OF THE LITERATURE
C. Garcia 1 , D. Constenla 1 1 International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, USA Background: To review the economic evidence of hospitalized invasive pneumococcal disease (IPD) among age groups ≥5 years in the Latin American and Caribbean (LAC) region. Methods: We systematically searched 5 databases of published and unpublished literature to identify studies from 1980-2012 presenting economic data for IPD cases among persons ≥5 in the LAC region. We summarized data from studies we found by study characteristics (e.g. study location, age groups), methodology (e.g. time horizon, perspective), and economic outcome data (e.g. cost per case, life years gained). Authors of unpublished studies were contacted for additional information on study methods and outcomes. Results: A total of 15 studies presenting economic data on IPD from 8 LAC countries were identified. Studies were exclusively among older adults (69% among ≥50 years old; 44% among ≥60 years old). Most studies were from the health system or third party payer perspective and included only direct medical and non-medical costs. Nine studies assessed the costs and benefits of pneumococcal vaccination and 6 assessed the total cost and cost per case of pneumococcal disease. The cost per single case of meningitis was found to be greater than for pneumonia (average, US$17,114 versus US$4,817). The cost of meningitis sequelae was found to be substantial (19-75% of acute management costs in two studies). Conclusion: Economic data on IPD among ≥ 5 year olds and hospitalized IPD among this population are limited. Available studies indicate that IPD incurs significant costs to society.

BUDGET OPTIMIZATION MODEL FOR PNEUMOCOCCAL VACCINATION IN INFANTS AND ELDERLY: THE CASE OF SPAIN AND THE NETHERLANDS
E. Delgleize 1 , B. Hoet 2 , B. Standaert 1 1 Health Economics, GlaxoSmithKline Vaccines, Wavre, Belgium; 2 Medical Affairs, GlaxoSmithKline Vaccines, Wavre, Belgium Background and Aims: Recommending bodies today may have to consider routine immunization with pneumococcal conjugate vaccines (PCV) in two groups (infants and elderly). This analysis aimed to identify an optimal PCV strategy within a constrained budget from the health-care payer perspective. Two countries, Spain and The Netherlands with different baseline PCV uptake, pediatric schedules and age indication in elderly are considered in this analysis.

Methods:
We developed an optimization model that is linked to a prevalence-based disease management submodel. This program allows to find an optimal solution given an objective function (minimize cases; minimize quality-adjusted life-years lost; minimize life-years lost) under budget constraints. In case of budget increase, the model calculates an optimal vaccine uptake in both groups. The vaccine efficacy against overall community-acquired pneumonia (VE-CAP) in elderly needed to justify a switch from infant to elderly vaccination is also estimated. Herd protection resulting from infant vaccination is included and varies with VE and uptake. Results: In Spain, though pneumonia disease burden is high in the elderly, the model estimates that additional budget should be first allocated to increase uptake amongst infants, irrespective of VE-CAP in elderly. In The Netherlands, the VE-CAP in the elderly would have to be at least 3-fold higher than that of infants to prioritize elderly vaccination. Conclusion: VE-CAP in the elderly would have to be very high to prioritize elderly vaccination. Infant immunization with PCV is identified as the optimal strategy to reduce the impact on invasive and non-invasive pneumococcal disease in both countries. Background: Few have assessed the global pneumococcal disease burden in persons ≥5 years. We assessed available data to estimate one component, pneumococcal meningitis, by geographic region pre-pneumococcal conjugate vaccine (PCV) introduction. Methods: We systematically searched literature and national surveillance databases from 1980-2010 for agestratified (5-19, 20-64, ≥65 years) pneumococcal meningitis incidence and inputs for two indirect estimation approaches: invasive pneumococcal disease (IPD) incidence multiplied by proportion of IPD that is meningitis and all-cause bacterial (AB) meningitis incidence multiplied by proportion of AB meningitis due to pneumococcus for persons >5 years. Results: Of 21,331 studies, 21 had pneumococcal meningitis incidence data. Data were sparse in Africa, Asia, and LAC (Table). For Asia, none of the approaches had data in every age strata.
Conclusion: Data to estimate pneumococcal meningitis incidence in persons ≥5 years were limited from Africa, Asia, and LAC, especially from adults ≥20 years. At least one approach is possible in all regions. For Asia, it requires extrapolating from the age distribution pattern from other regions. Background: Few have assessed the global pneumococcal disease burden in persons ≥5 years. We assessed data available to estimate one component, pneumococcal pneumonia, by geographic region pre-pneumococcal conjugate vaccine (PCV) introduction. Methods: We systematically searched studies conducted 1980-2010 for data to estimate regional pneumococcal community-acquired pneumonia (PCAP) incidence in persons >5 years. PCAP can be estimated directly, or indirectly by summing pneumococcal bacteremic pneumonia (PBP) and non-bacteremic pneumonia (non-PBP) incidence. The ratio PBP:non-PBP (previously estimated) can be applied to PBP incidence to estimate CAP incidence. Inputs to estimate PCAP indirectly are listed in the Table. Models using multiple inputs require age-stratified (5-19, 20-64, ≥65 years) data. Results: PCAP incidence data was only available for North America and Oceania. For other regions, no single study had data from all age-strata, and no indirect approach had data for all age-strata.
Conclusion: Data to estimate pneumococcal pneumonia incidence globally were lacking in adults. No single method was possible for all regions, but each region has data for ≥1 method. Background and Aims: Global pneumococcal disease burden in persons ≥5 years has not been described. We aimed to estimate one component, pneumococcal meningitis, by geographic region pre-pneumococcal conjugate vaccine (PCV) introduction. Methods: A systematic review of 17 literature databases and national surveillance reports identified studies conducted 1980-2010 reporting pneumococcal meningitis incidence in persons ≥5 years pre-PCV introduction.

ISPPD-0511 Global Pneumococcal Disease and Policies for Control
Regional meta-estimates of age-stratified (5-19, 20-64, ≥65 years) incidence were obtained for studies reporting standard errors. Year 2013 UN Region age-specific population sizes were used to weight these and estimate incidence and cases ≥5 years. Results: All regions had ≥1 study representing all age strata except Africa (but all age-groups were represented) and Asia (no data). Adults ≥65 years had highest incidence in all regions. Cases were CSF culture-positive or latex agglutination-positive (occasionally blood culture-positive); PCR testing was rare.  Background and Aims: Afghanistan is said to have high case fatality rate of acute respiratory infections (ARI) among under-5 year old children, however no data have been published. Serotype distribution of pneumococcus is unknown. We aimed to determine case fatality rate, risk factors for death, and prevalence of pneumococcus with its serotype distribution in Afghanistan. Methods: A prospective study was conducted in the Paediatric Ward of Mazar-e-Sharif Regional Hospital, Afghanistan from December 2012 to March 2013. Under-5 year old children admitted with clinical pneumonia (WHO) were recruited. Nasopharyngeal samples were tested for Streptococcus pneumoniae (pneumococcus), Haemophilus influenzae and Moraxella catarrhalis by multiplex PCR. Molecular serotyping and quantification of pneumococcus was done by nanofluidic real time PCR. Results: Total recruited children were 639 and 326 samples were collected. Case fatality rate of pneumonia was 12.1% (75/639). Risk factors associated with death were: age less than 1 month, unable to breastfeed, chest indrawing, cyanosis, altered consciousness, hypothermia, hypoxemia and acute malnutrition. S. pneumoniae, H. influenzae and M. catarrhalis were detected in 41.1%, 31.2%, and 22.7% respectively. Twenty-two pneumococcal serotypes/serogroups were identified. PCV13 covered 38.1% of prevalent serotypes. Multiple serotypes were present in 15.9% (21/132) of pneumococcus colonized children, and their prevalence was higher, 36.3% (4/11) in death cases than 12.7% (15/118) in survived cases (p = 0.057). Pneumococcal load was also higher in death cases (4.6 log10/mL) than survived cases (3.1 log10/mL) (p = 0.07). Conclusion: High case fatality rate of ARI should be addressed. Wide range of pneumococcal serotype distribution and low serotype coverage of PCV13 warrant further studies. Background: Pneumococcal conjugate vaccines (PCV) available in New Zealand (NZ) and listed on National Immunisation Schedule for the prevention of pneumococcal disease (PD) are (i) 10-valent PCV (PCV10) for all infants and children, and (ii) 13-valent PCV (PCV13) for high risk children <5 years and those aged <16 years preor post-splenectomy or with functional asplenia. The objective of this economic analysis was to assess the costeffectiveness of PCV13 replacing PCV10 in NZ based on 3+1 dosing for all infants and children. Methods: A 1-year cross-sectional, steady-state, population model was developed from a payer's perspective to estimate PD cases avoided (bacteraemia, meningitis, inpatient and outpatient pneumonia, simple and complex otitis media). Recent disease incidence, serotype coverage, and population data obtained from the NZ Ministry of Health (MoH). Utilities sourced from published literature. Medical costs derived from the MoH and literature. Price parity per dose was assumed. Vaccine direct effectiveness derived from 7-valent PCV, and observed effectiveness of PCV10 and PCV13. The base analysis also included PCV13 indirect effects. Results: PCV13 for primary immunisation reduces the number of PD cases (n = 4,026) relative to PCV10. It also leads to an increase in the number of life-years (n = 352) and quality-adjusted life-years saved (n=387). Annual direct medical costs (including vaccination) were estimated to decrease by more than $5.7 million when vaccinating with PCV13. Thus, PCV13 dominated. Sensitivity analyses supported this in all scenarios. Conclusion: Sole supply of PCV13 for primary immunisation against PD in NZ is a cost-saving programme compared with PCV10. Introduction: Cost-effectiveness analyses (CEAs) of 10-valent and 13-valent pneumococcal conjugate vaccines (PCV10, PCV13) have been based on extrapolation of efficacy and effectiveness data from the 7-valent vaccine (PCV7). PCV10/PCV13 efficacy and effectiveness data are now public, therefore we assessed application of this data into CEAs. Methods: Publicly available PCV10 and PCV13 efficacy and effectiveness data through October 2013 were analyzed. Model inputs recommendations for both vaccines were based on assessment of competent and reliable scientific evidence, statistical significance, and across-study comparability. Results: No head to head vaccine efficacy or effectiveness data are available. Data demonstrate higher-valent vaccines reduced IPD, pneumonia, and otitis media in vaccinated individuals, supporting direct protection. Otitis media and pneumonia endpoints differ across studies, limiting ability to directly compare impact across vaccines. PCV13 consistently exhibited rapid and robust indirect effects across age groups, while one PCV10 study showed reduction in unvaccinated children <5 years. PCV13 consistently demonstrated significant 19A protection, while PCV10 has not. Neither PCV10, nor PCV13's 6 additional serotypes have yet demonstrated significant evidence of effect on disease or carriage caused by non-vaccine specific serotypes, non-typeable Haemophilus influenzae, or other pathogens (Table).

Conclusion:
Data informing higher-valent PCV CEAs continue to evolve. Model estimations of direct protection from pneumococcal diseases should be based on a common value, adjusted to reflect local coverage of serotypes included in the vaccines. PCV13 indirect effects should be included; PCV10 indirect effects should be limited to sensitivity analyses until proven.

Methods:
We developed a Markov model simulating the lifetime evolution of a local birth cohort without vaccination as a base case. Different scenarios were considered for comparison, according to the evidence of effectiveness, efficacy and herd effect of PCV. We established a universal 3 (2+1) dose vaccination schedules and we modeled the burden of diseases from local and regional data and paraguayan cost. The results were expressed as incremental cost in 2013 US dollars per life years gained (LYG) Results: The incremental cost per LYG for PCV13 (2+1) according to PCV-13 pneumonia hospitalization effectiveness regional data with and without herd immunity was 618, 96 USD and1.095 USD, respectively. The incremental cost per LYG for PCV13 (2+1) according to non-inferiority PCV-7 efficacy and herd effect, was 813, 17 USD. The incremental cost per LYG for PCV10 (2+1) according to efficacy/effectiveness data was 1.219, 74 USD. The model was sensitive to variation with regards to assumptions around pneumonia incidence and mortality rate. Conclusion: All scenarios are cost effective and lower than Paraguay gross domestic product. Evidence of herd effect associated with PCV13 improves pneumococcal vaccines and is cost saving in Paraguay.

SEROTYPE PREVALENCE AND ANTIBIOTIC RESISTANCE IN STREPTOCOCCUS PNEUMONIAE CLINICAL ISOLATES IN SFAX,SOUTH OF TUNISIA
A. Hammami 1 , I. Kotti-jmal 1 , F. Mahjoubi-Rhimi 1 , A. Znazen 1 , S. Mezghani-Maalej 1 , B. Mnif-Chaaben 1 , S. Ktari-Chaari 1 1 Microbiology, Habib Bourguiba hospital, Sfax, Tunisia Background and Aims: Streptococcus pneumoniae causes a wide range of infections that could be deliterious. Prevention through vaccination is a valuable tool to decrease the burden of disease. Nevertheless, none of the marketed internationally vaccines, PCV 7, PCV 10 and PCV 13 is currently part of the national program of immunization in Tunisia. We undertook this study to determine the serotype distribution and to analyze the antimicrobial resistance of S. pneumoniae isolates . Methods: All pneumococcal strains isolated in the microbiology laboratory of the University Hospital, Sfax, Tunisia, from January 2012 to August 2013 were included. S. pneumoniae was identified by Gram straining, optochin susceptibility and bile solubility. Antimicrobial susceptibility was determined by the disk diffusion and E test methods. Serotyping was performed by multiplex PCR. Statistical analysis was done using SPSS 2O. Results: Among 125 collected pneumococcal isolates, 39 were invasive isolates (31%). The mean age of patients was 29.7 years. Seventy four percent of the strains were penicillin non-susceptible (PNSP). Forty four percent had decreased susceptibility to amoxicillin and 23.2% to cefotaxim. The PNSP were more frequently resistant to other antibiotics. Serotype 14 was the most frequently isolated (21.4%) followed by serotypes 19F (20.4%), 6A/6B (10.2%), 23F (9%) and 3(8.2%). Serotype 19F was associated with higher level of PNSP (p = 0.03). The potential coverage by the 7, 10 and 13 valent pneumococcal conjugate vaccines were 65.3%, 65.3% and 77.6% respectively. Conclusion: A high rate of S. pneumoniae antibiotic resistance is observed in Tunisia. Conjugate vaccines and particularly PCV 13 provide good coverage for pneumococcal isolates . Introduction: Pneumococcal infections are vaccine-preventable diseases. Immunization against pneumococci is recommended in Poland but not reimbursed. The aim was to assess the coverage of the vaccination. Material and Methods: A retrospective chart analysis of 1356 children in a large primary health care centre. The type of the vaccine, the number of doses, compliance with the dosing regimen and age of first doses were analyzed. Results: Pneumococcal conjugate vaccine was administered in 499 children (36.8%), in majority of them (230/499; 46.1%) the immunization started in the first 6 months of life, in 63/499 patients (12.6%) at age 7-11 months, in 63/499 children (12.6%) at age 12-23 months and in 143 patients (28.7%) at age over 24 months. The schedule was correct only in 16/230 children immunized in first 6 months of life, 4/63 children at age 7-11 months and 54/143 children immunized at age 12-23 month. Conclusion: Pneumococcal immunizations are used ineffectively in Polish children. Introduction of reimbursed, universal immunization would improve the coverage and adherence to recommended schedule. Background: Pneumococcal-conjugate vaccines (PCVs) represent the most promising pneumococcal immunization approach for the younger children. The herd immunity and overall beneficial protections will only be pronounced if this vaccine is to be implemented into the national childhood immunization scheme. Thus, the pneumococcal serotype distribution data is important for countries like Malaysia at which none of the PCV has been widely in-use to evaluate the potential coverage of PCVs. Methods: A total of 208 pneumococcal isolates from various sites stored at the microbiology laboratory of University of Malaya Medical Centre, Kuala Lumpur from 1997 -2012 were subjected to multiplex PCR serotyping. Minimum inhibitory concentration (MIC) was determined using agar dilution method or broth microdilution method. Statistical analysis was performed using chi-squared or fisher's exact tests whenever appropriate. Results: The predominant serotype was 19F (33%), followed by 23F (9%), 6A/B (6%), 1 (6%), and 14 (5%). Penicillin non-susceptible strain (PNSP) constituted 51% of all isolates. Statistical testing found serotype 19F to be associated with penicillin resistance (62%). Moreover, 19F was associated with noninvasive site (44%) while serotype 19A was associated with invasive site (11%). Notably, the invasive isolates were associated with penicillin susceptible strains while the noninvasive isolates were likely to be PNSP. Conclusion: The PCVs (PCV7, PCV10, and PCV13) are expected to cover substantial portions of 56%, 63%, and 71% of Malaysian serotypes, respectively. Hence, the use of PCVs would benefit the local population. Furthermore, continuous surveillance to monitor the temporal fluctuation in the local serotypes is highly desired. Background: Pneumococcal serotyping in Alberta is centralized at the Public Health Laboratory located in Edmonton, Alberta. PCV7 was introduced in Alberta in 2002 and PCV13 in 2012. We report the serotype distribution for cases of invasive pneumococcal disease from January 2003 to August 2014. Methods: Invasive pneumococcal disease (IPD) is a notifiable disease reportable to Provincial Health Authorities. This results in IPD isolates in Alberta (population 3.5 million) forwarded from diagnostic microbiology laboratories to the Public Health Laboratory for serotyping. This serotyping data from January 2003 to August 2013 was collated for this analysis. Results: From 2003 to August 2013, the top ten serotypes in Alberta in order were serotype 5 (527 isolates -10.9%), 8 (363-7.5%), 22F (354-7.3%), 4 (340-7%), 3 (321-6.6%), 19A (317-6.5%), 7F (241-5.0%), 14 (205-4.2%), 19F (135-2.8%) and 12F (134-2.8%). This large collection of serotype 5 isolates was the result of a previously documented outbreak that occurred from 2005 to 2008. In 2012, the last full calendar year collected, the top ten serotypes in order were 22F (42-11.1%), 19A (40-10.6%), 7F (34-9.0%), 20 (31-8.2%), 8 (29-7.7%), 4 (26-6.9%), 3 (24-6.3%), 23B (15-4.0%), 33F (13-3.4%), and 15A (13-3.4%). The only PCV7 serotype remaining in the top ten serotypes is serotype 4. All other PCV7 serotypes have declined to levels of less than 1%. Conclusion: The serotype distribution of pneumococcal isolates from cases of IPD in Alberta, Canada has fluctuated over the 12 years. Of particular concern is the increase in nonPCV13 serotypes, 22F, 20 and 8.

ISPPD-0305 Global Pneumococcal Disease and Policies for Control
Objectives: 10-valent pneumococcal conjugate vaccine (PCV10) was introduced into the immunization routine program of Brazil in 2010, with three schedules: 3+1, 2+1, and single dose (≥12 mo). This study aimed to investigate coverage and compliance six to eight months after the introduction of PCV10 into childhood vaccination at Goiania (~1,300,000 inhabs), a municipality where both routine immunization and vaccination campaign have been used. Methods: A household survey was conducted from December 2010 through February 2011. A systematic sampling was used to enroll 1,237 children aged 7-11 months (n = 647), and 15-18 months (n = 590). Vaccination status was retrieved from immunization cards. A questionnaire on socioeconomic characteristics was applied to assess variables associated with vaccination coverage (number of children who received all recommended doses) and compliance (number of children who received all doses without any delay). Results: Vaccination coverage was 53.4% (95%CI 50.7-56.2%), ranging from 88.3% (≥12 mo) to 39.3% (7-11 mo). Compliance with the recommended schedule was 16.6% (95%CI 14.6-18.7%), with variations of 18.8% (≤6 mo), 6.0% (7-11 mo), and 35.6% (≥12 mo). Variables independently associated with not being completely vaccinated (coverage) were mother's lower schooling (odds Ratio/OR=1.68; 95%CI 1.08-2.61), and ≥3 household children (OR=2.06; 95%CI 1.09-3.91). Variables independently associated with noncompliance were mother's lower schooling (OR=1.67; 95%CI 1.12-2.50), and not having private health insurance (OR=1.46; 95%CI 1.03-2.07). Conclusion: Immunization program must focus on initiatives that might increase 'complete and on time' vaccination rates, especially for children in catch-up schedules and those in the lowest socioeconomic strata.  Background and Aims: Acute otitis media (AOM) is the most common cause of pediatric outpatient visits. Streptococcus pneumoniae is one of the leading bacterial pathogens causing AOM. We assessed the impact of PCV10 in outpatient visits for all-cause otitis in children.

Methods:
We used individual-level secondary data from outpatient Electronic Medical Records of patients covered by the Public healthcare system in Goiania, Brazil, which covers 63% of all pediatric visits. Study period was January 2008-August 2013. All children aged 2-23 months with diagnosis of all-cause otitis (ICD-10 codes: H65-H67) were identified. Rates of overall and all-cause otitis outpatient visits per 10,000 children were calculated, considering monthly population estimates from 2000 and 2010 census. PCV10 was introduced in June 2010, being the intervention evaluated. Rates of outpatient visits due to other causes, except respiratory, ear and mastoid diseases, were the comparator. A time-series analysis was conducted fitting a Prais-Winsten autoregressive model, adjusted for monthly, random and seasonal variations. Results: During the study period 456,153 outpatient visits among children aged 2-23mo were identified, of which 6,177 (1.4%) were due to all-cause otitis, and 241,379 (52.9%) to other causes. Time-series analysis indicated a monthly average rate reduction for all-cause otitis of 1.58% (95%CI: 0.41%-2.73%; p=0.009) after vaccination ( Figure  1), whereas the monthly rate of outpatient visits due to other causes increased by 0.51% (95%CI: 0.06%-0.95%; p=0.026). Conclusion: Three years after the introduction of PCV10 in Goiania, a significant reduction in rates of outpatient visits due to all-cause otitis was observed in children aged 2-23 months.  Background and Aim: We investigated the impact of otitis media (OM) and its sequelae on hearing impairment in Indonesian school children. Methods: This was a prospective epidemiological survey in a sample of 7005 public school children (6-15 years) from urban and rural subdistricts, in 6 locations on 5 islands in Indonesia. Children had otoscopic and hearing screening tests. Those with abnormalities on either, had diagnostic audiometry and tympanometry performed. Results: The overall rate of hearing impairment (HI) in the school children was 167/10,000. OM was detected in 172 children (2.5%), consisting of acute otitis media (AOM) (17%), otitis media with effusion (OME) (15%) and chronic suppurative otitis media (CSOM) (67%). Among children with ear disease, 41% had mild and moderate conductive HI, mostly bilateral. OME resulted mostly in mild HI(9/11; 81%), while CSOM resulted mostly moderate HI(26/41;63%), a significant difference (P= 0.01). HI was disabling in 31/56 children, mostly (26/31) due to CSOM. Of the better ear hearing there was a significantly higher rate of OM related HI in rural (113.3/10,000), than in urban areas (48/10,000), p = 0.002. In rural areas this was mostly due to CSOM 96/10,000 whereas in urban areas AOM and OME contributed to about ½ of the HI. Conclusion: Otitis media or its sequelae contributed to the cause of almost two thirds of all HI in Indonesian school children. CSOM was associated with about 50% of the HI in urban children and 85% in rural school children. Prevention of CSOM could potentially prevent a significant burden of HI in Indonesian school children. Dept. of Surgery, University of Nairobi Kenyatta National Hospital, Nairobi, Kenya; 2 Dept. of Global Health, University of Colorado at Denver, Denver, USA Background: We investigated the impact of otitis media (OM) and its sequelae on hearing impairment in Kenyan children.

ISPPD-0500 Global Pneumococcal Disease and Policies for Control
Methods: This was a prospective study of 13,109 school and preschool children aged 2-15 years in 9 regions in Kenya. All children had a history, otoscopic examination and a hearing screen test. Children with otoscopic abnormalities or who failed the hearing screen underwent diagnostic audiometry and tympanometry. Results: The prevalence of hearing impairment (HI) was 1032/10,000, with no difference between rural and urban areas. OM was detected in 487 children (3.7%), acute otitis media (AOM) (80, 16%), otitis media with effusion (OME) (189, 39%), Retractions (15, 3%), and chronic suppurative otitis media (CSOM) (203, 42%). Among children with OM, 61% had mild and moderate conductive HI, mostly bilateral. OME accounted for most of the mild HI, while CSOM accounted for most of the moderate HI, = 0.0024. In 51 cases (0.4%), the HI was disabling, 69% due to CSOM. Of the worse ear hearing levels, the rates/10,000 urban and rural children for HI due to AOM were 24 and 36; OME were 116 and 101; and for CSOM were 161 and 133 respectively. The rate of disabling HI associated with CSOM in urban and rural children were 113/10,000 and 84/10,000 respectively. There was an increasing incidence of disabling HI in those with CSOM from 35% in those <6 to 65% in those 14 years of age (p = 0.028). Conclusion: CSOM results in a significant burden of hearing impairment in Kenyan school children, and its early prevention might mitigate some of this burden.  Background: Pneumococcal conjugate vaccine (PCV) is not included in India's Universal Immunization Program (UIP) but is available on the private market and is recommended by the Indian Academy of Pediatrics (IAP). Methods: A random selection of IAP member pediatricians was surveyed regarding their practices for administering PCV. Half were assigned to complete the survey over the telephone with a study coordinator and half were emailed the survey to complete at their discretion. Sample size was calculated to give proportions with ±5% confidence intervals, assuming a 50% response rate. Results: Response rates differed significantly, with 59/382 (15%) completing the survey by email and 275/382 (72%) by telephone, for an overall response rate of 44%. 73% of pediatricians administer PCV to some patients, but only 7% administer to all patients. The most common reason for not prescribing PCV was that it is too expensive (94%). Other reasons included: unsure of vaccine efficacy (10%), unsure risk of disease warrants it (5%), unsure of vaccine safety (2%), concerns about serotype coverage (1%). Results differed by type of practice. Pediatricians at government health centers were the least likely to administer PCV. 88% of pediatricians have their own children vaccinated or recommend PCV to close friends, and 83% would like to see it included in the UIP. Conclusion: There is support for the use of PCV by IAP member pediatricians, with a majority already prescribing the vaccine to some patients. High cost is the most commonly reported barrier to universal access. For each diagnostic category we estimated the societal costs to the health system (patient specific costs and bed day costs) and out-of-pocket costs (OOP) before, during, and one week after discharge from hospital/OP visit. Results: A total of 340 children were enrolled; 29 meningitis, 36 sepsis, 175 inpatient and 100 outpatient pneumonia (21, 32, 94 and 50 respectively from rural Gambia). Mean provider costs per case for treating OP, IP, pneumococcal sepsis and meningitis were US$8, US$64, US$87 and US$124 respectively. The average family's OOP per case for OP, IP, sepsis, and meningitis in the rural area were US$2, US$20, US$44 and US$31 respectively, and US$10, US$44 US$41 and US$60 in the urban setting. The economic burden of inpatient pneumonia, pneumococcal sepsis and meningitis increased to US$722, US$407 and US$688 respectively when time loss from work by family was taken into account. Conclusion: The economic burden of pneumococcal disease in The Gambia is substantial, with high disease annual incidence before PCV introduction of 500/100,000 in infants and 200/100,000 in children under 5 years. Prevention by PCV promises to be cost-effective and potentially cost-saving.

SEROTYPE DISTRIBUTION, ANTIMICROBIAL RESISTANCE AND EXPECTED VACCINE COVERAGE FOR INVASIVE PNEUMOCOCCAL DISEASE IN INDIAN ADULTS
B. Veeraraghavan 1 , R. Jayaraman 1 , R. Verghese 1 , K. Thomas 2 1 Microbiology, Christian Medical College, Vellore, India; 2 General Medicine, Christian Medical College, Vellore, India Background and Aim: Streptococcus pneumoniae infections in adults are associated with substantial morbidity, mortality, and costs. Pneumococcal infections in adults for the most part underestimated especially in low and middle income countries. Pneumococcal polysaccharide vaccines (PPV-23) are available to prevent penumococcal infections. Despite its controversies it is in routine use for adults in most developed countries. Developing countries will need pneumococcal disease burden and serotype prevalence data in adults for implementing PPV-23 for routine use. Here we present the pneumococcal serotype data for adults generated over a period of two decades and aim to estimate the expected protective coverage of available vaccines Materials and Methods: Invasive S. pneumoniae isolates from more than six years of age group was included in this study. Isolates were serotyped with co-agglutination technique. Antibiotic susceptibility profile for the isolates were determined by vitek system 2 Results: The most common serogroup/types in India causing invasive bacterial infections in adults are 1, 19, 6, 3 and 5. 93% of the isolates were non-susceptible to co-trimoxazole, and 20% of the isolates were non-susceptible to erythromycin. 8% of the isolates were non-susceptible to penicillin and only 4% of the isolates were non-susceptible to cefotaxime Conclusion: PCV-13 vaccine can protect 62% of the isolates causing invasive pneumococcal infections. PPV-23 can provide better coverage with 72% coverage for IPD in Indian adults. Routine use of PPV-23 can reduce a substantial limit of pneumococcal diseases in Indian adults  (2011:162;2012:92). 33 serotypes were identified from IPD. PCV13 serotypes declined from 87.9% to 73.9% (p < 0.001). Serotype 14 reduced from 24.1% to 13.7% (p = 0.01) and not PCV13 serotypes increased from 12.1% to 26.1% (p < 0.001).

Conclusion:
We observed a decline in clinical pneumonia, pneumococcal pneumonia and meningitis in children under 24 months, pneumonia showed greater reduction. PCV13 serotypes were reduced (mainly serotype 14) and not PCV13 serotypes which increased. These preliminary data is from National Epidemiological Surveillance and should strengthen the need to monitor the strategy. Introduction: Argentina introduced PCV13 into National Immunization Programme in 2012 for healthy children <1 year with 2 + 1 schedule and catch-up between 12-24 months. Sentinel pneumococcal surveillance units (SU) started in the same year in 7 hospitals to assess the impact of this strategy. Objectives: Describe the evolution of bacterial pneumonia hospitalizations pre and post PCV13 introduction and describe confirmed cases of Streptococcus pneumoniae isolated after vaccine introduction. Methods: Retrospective (2008Retrospective ( -2011Retrospective ( ) and prospective (2012Retrospective ( -2013 data of probably bacterial pneumonia under 5 years according to clinical and XRay were analyzed in 2 of 7 SU. Time-series were performed calculating moving averages (range= 4). Invasive pneumococcal disease (IPD) prospective data (2012-2013) were analyzed in 3 SU. Results: There was a 41.3% reduction (95% CI 15.7-66.9%, p <0.001) of monthly admissions for pneumonia.

Serotype
History of immunization with PCV13 Total 1 dose 2 doses 3 doses Conclusion: There was a significant reduction in hospitalizations for pneumonia. To assess impact of the vaccine, it is necessary to continue with epidemiological and laboratory surveillance in children.
No conflict of interest