Impact of Widespread Introduction of Pneumococcal Conjugate Vaccines on Pneumococcal and Nonpneumococcal Otitis Media

Transcription

1 Clinical Infectious Diseases MAJOR ARTICLE Impact of Widespread Introduction of Pneumococcal Conjugate Vaccines on Pneumococcal and Nonpneumococcal Otitis Media Shalom Ben-Shimol, 1,2 Noga Givon-Lavi, 1,2 Eugene Leibovitz, 1,2 Simon Raiz, 2,3 David Greenberg, 1,2 and Ron Dagan 2 1 Pediatric Infectious Disease Unit, Soroka University Medical Center, 2 Faculty of Health Sciences, Ben-Gurion University of the Negev, and 3 Department of Otolaryngology, Soroka University Medical Center, Beer-Sheva, Israel Background. Pneumococcal conjugated vaccines (PCVs) impact on complex otitis media (OM; including recurrent, nonresponsive, and chronic OM with effusion) was greater than that on simple, acute OM in previous studies. Since complex OM is often a polymicrobial disease, we speculated that reduction of complex OM by PCVs would be associated with reduction of nonpneumococcal OM. Methods. In a prospective, population-based, active surveillance, all OM episodes submitted for middle ear fluid culture in children <3 years from 2004 through 2015 were included. Three sub-periods were established: pre-pcv, PCV7, and PCV13. Incidence rate ratios (IRRs) comparing the 3 periods were calculated for pneumococcal, nontypable Haemophilus influenzae (NTHi), Moraxella catarrhalis, Streptococcus pyogenes, and culture-negative OM. Results. Overall, 7475 episodes were included. Of all-nthi cases in the pre-pcv period, 34% were mixed with Streptococcus pneumoniae. IRRs (95% confidence interval) comparing the pre-pcv to the PCV13 period were 0.02 ( ), 0.12 ( ), and 0.18 ( ) for PCV7+6A serotypes, 5 additional PCV13 serotypes, and all-pneumococcal OM, respectively; non-pcv13 serotype episodes were not significantly reduced. IRRs for single NTHi, mixed NTHi + S. pneumoniae, and all-nthi OM were 0.30 ( ), 0.18 ( ), and 0.25 ( ), respectively. Moraxella catarrhalis, S. pyogenes, and culture-negative episodes were also significantly reduced. Conclusions. Both pneumococcal and non-pneumococcal OM episodes, enriched with complex cases, declined substantially in children <3 years following sequential PCV7/PCV13 introduction. The reduction in non-pneumococcal episodes may be attributed to early OM episodes prevention, resulting in a lower rate of complex, often non-pneumococcal OM. Keywords. nontypable Haemophilus influenzae; complicated otitis media; pneumococcal conjugate vaccine; children; surveillance. Otitis media (OM) is a multifactorial, often polymicrobial disease. Clinical manifestations range from early, uncomplicated, sometimes asymptomatic OM (simple OM) to complex OM such as recurrent, nonresponsive, spontaneously draining and chronic OM with effusion [1]. The main pathogens are Streptococcus pneumoniae and nontypable Haemophilus influenzae (NTHi). Streptococcus pneumoniae is the leading pathogen in early OM (usually at age <6 months) [2]. Serotypes included in the 13-valent pneumococcal conjugate vaccine (PCV13) are more common in early onset (simple) OM than in complex OM [1]. Middle ear damage from early events may result in increasingly complex, often polymicrobial OM episodes [1, 3], with a diminishing relative role of S. pneumoniae (especially Received 16 February 2016; accepted 17 May 2016; published online 25 May Correspondence: R. Dagan, Pediatric Infectious Disease Unit, Soroka University Medical Center, Beer-Sheva, Israel (rdagan@bgu.ac.il). Clinical Infectious Diseases 2016;63(5):611 8 The Author Published by Oxford University Press for the Infectious Diseases Society of America. All rights reserved. For permissions, journals.permissions@oup.com. DOI: /cid/ciw347 PCV13 serotypes) [4 6] and an increasing NTHi role, often in association with mixed infections and biofilm formation [7 11]. NTHi is also a major pathogen in other recurrent, nonresponsive and chronic respiratory infections [12 15]. Thus, it is plausible that early, acute pneumococcal OM predisposes to complex OM, often involving nonvaccine pneumococci and pathogens other than S. pneumoniae (mainly NTHi) [1, 16]. In this case, widespread use of PCV may prevent this progression. In southern Israel, most children are born at the only medical center in the region (Soroka University Medical Center [SUMC]), where they also receive medical services. Additionally, >95% of middle ear fluid (MEF) cultures from the region are sent to the SUMC laboratory [17]. This enabled us to conduct a prospective, population-based study to describe pathogenspecific incidence of OM necessitating MEF culture. The sequential 7-valent pneumococcal conjugated vaccine (PCV7)/PCV13 was introduced to the Israeli National Immunization Program (NIP). We recently reported on the early impact of PCV7/PCV13 on the incidence of pneumococcal OM episodes necessitating MEF culture in children aged <2 years PCV and Nonpneumococcal Otitis Media CID 2016:63 (1 September) 611

2 [17]. Remarkable reductions of all-pneumococcal and all-cause OM were observed, with near elimination of PCV13 serotypes (VT13) within 5 years of PCV introduction. Using a similar design, we expanded the study to determine the impact of PCV on OM caused by nonpneumococcal otopathogens. In addition, we expanded the age group to children aged <3 years and added 2 years of surveillance. We hypothesized that widespread PCV use will impact pathogens beyond VT13, by preventing early pneumococcal OM. Moreover, we hypothesized that the impact on pneumococcal OM will be greater and faster than the impact on other pathogens and will be observed earlier in children aged 4 23 months compared with children of otherages(duringtheearlypost-vaccination period, older children could have suffered from early pneumococcal OM before PCV introduction, and children aged <4 months [who received 1 PCV dose] are mainly dependent on indirect protection through reduced exposure, expected only later after NIP initiation). Specifically, our objective was to document the pathogenspecific incidence dynamics of OM necessitating MEF culture in children aged <3 years 5 years before and 6 years after PCV introduction. MATERIALS AND METHODS This ongoing, prospective, population-based, active surveillance spanned an 11-year period, July 2004 through June The SUMC Ethics Committee approved the study. Detailed methods have been described previously [17] and are presented here in brief. Setting and Study Population In southern Israel (the Negev region), Jewish and Bedouin populations live side by side. Although the socioeconomic conditions and lifestyles of the 2 groups differ, both have access to the same medical services. The Jewish population, which is mainly urban, resembles developed populations. The Bedouin population, formerly desert nomads, while being in transition to a Western lifestyle, resembles developing populations, with high occurrence of infectious diseases including pneumococcal disease and complex OM [3, 7, 16, 17]. During the study period, the average total annual births were approximately (approximately 50% Bedouin); >95% of the Negev s children are born and receive medical services at the SUMC. The study population included children aged <3 years with OM episodes resulting in MEF culture. Indications for referral, tympanocentesis, or MEF culture were similar during the 11 years of study. These include (but are not limited to) young age, previous OM or tube insertion, high grade fever or toxic appearance, and spontaneous drainage. OM diagnosis was made by a pediatrician, family physician, or otolaryngologist, as previously described; cultures were obtained by tympanocentesis or pus collected from draining ears (of <7 day duration) [17]. Demographic and clinical information was obtained from medical charts, the child s physician, or parents, as appropriate for children with positive cultures, and included culture date, age, ethnicity, previous OM episodes, and recent antibiotic treatment [17]. Bacteriology Specimen swabs were placed in transport medium and processed as previously described [17]. Streptococcus pneumoniae, NTHi, Moraxella catarrhalis, andstreptococcus pyogenes were defined as OM pathogens and identified as previously described [18]. Pneumococcal serotyping was performed with the Quellung reaction using antisera from Statens Serum Institut, Copenhagen, Denmark. PCV7/PCV13 Vaccine Uptake PCV7 NIP was initiated in July 2009 with a catch-up campaign in children aged <2 years [17]. In November 2010, PCV13 replaced PCV7, without further catch-up. Vaccine uptake evaluation methods have been previously described [19]. By June 2011 and December 2012, approximately 80% of children aged 7 11 months received 2 PCV7 doses and approximately 90% received PCV13 doses; by June 2014, approximately 95% received 2 PCV13 doses. By June 2011 and December 2012, 36% of children aged months received 3 PCV7 doses and 87% received PCV13 doses; by June 2014, 91% received 3 PCV13 doses. Statistical Analyses OM episodes were separated by a 30-day interval for the same pathogen or any interval for different organisms. For pneumococcus, same organism was defined by identical serotype; for other pathogens, there was no further intraspecies characterization. Only 1 isolate was included per episode; if the same strain was isolated from both ears, only 1 was selected randomly [17]. Mixed infection was defined as simultaneous presence (in the same ear or both ears) of 2 defined pathogens. The following 3 subperiods were established in order to allow better impact appreciation: pre-pcv, July 2004 June 2008; PCV7, July 2009 June 2011; and PCV13, July 2013 June Mean incidences during PCV7 and PCV13 periods were compared with those from the pre-pcv period. The following 3 age groups were established: <12 (further divided to <4 and 4 11), months, and months. Incidences were calculated using the birth cohorts born in southern Israel, according to the Israeli Central Bureau of Statistics reports [20]. Incidence rate ratios (IRRs) and 95% confidence intervals were calculated for specific pathogens and overall OM and analyzed by ethnic groups (Jewish and Bedouin) and age groups. Data were analyzed using SPSS software. Univariate analyses were conducted using 2-tailed χ 2 test or Student t test, where appropriate. A P value <.05 was considered statistically significant. 612 CID 2016:63 (1 September) Ben-Shimol et al

3 RESULTS From July 2004 through June 2015, 7475 OM episodes with MEF cultures were recorded in children aged <3 years, of which 64% (4753) were culture-positive (Table 1). All pneumococcal, mixed pneumococcal + NTHi, single NTHi, and other (M. catarrhalis or S. pyogenes) culture-positive episodes constituted 2230 (30%), 902 (12%), 1969 (26%) and 404 (5%) of episodes, respectively. Children aged <4, 4 11, 12 23, and months comprised 937 (13%), 2929 (39%), 2816 (38%), and 793 (11%), respectively, of the study population. During the pre-pcv period, of children aged <4, 4 11, 12 23, and months, 60/161 (37%), 592/844 (70%), 607/717 (85%), and 122/144 (85%), respectively, had 1 risk factor for complex OM (antibiotic treatment in last month, >3 previous OM episodes, or previous tympanocentesis), by history. Pneumococcal OM OM caused by PCV7 serotype + serotype 6A (VT7 + 6A) and allpneumococcal OM rates declined by 65% and 30%, respectively, during the PCV7 period compared with the pre-pcv period (Table 2, Figure 1). In contrast, during this period, the 5 additional PCV13 serotype (VT5) OM rates increased nonsignificantly by 7%, and non-vt13 OM increased significantly by 48%. During the PCV13 period, overall pneumococcal and VT7 + 6A OM episodes declined by 75% and 94%, respectively, compared with the PCV7 period, with VT5 and non-vt13 episodes significantly reduced by 88% and 45%, respectively. The overall reduction rate from pre-pcv to PCV13 periods was 98%, 88%, and 82% for VT7 + 6A, VT5, and overall pneumococcal OM, respectively. A nonsignificant decrease in non-vt13 was observed during the PCV13 period compared with the pre-pcv period. Nonpneumococcal OM Overall, NTHi OM rates declined by 9% from the pre-pcv7 to the PCV7 period and by 72% from the PCV7 to the PCV13 period, resulting in an overall 75% decline (Table 2, Figure 1). Of all NTHi episodes, in the pre-pcv period, 34% were mixed, mostly with S. pneumoniae as dual pathogens (91%); the remainder involved M. catarrhalis, S. pyogenes, or both. Single- NTHi OM rates declined nonsignificantly by 7% from the pre-pcv to the PCV7 period. However, during the PCV13 period, a significant 70% decline from the pre-pcv period was observed. Mixed NTHi + pneumococcal OM rates declined by 20% from the pre-pcv to the PCV7 period and by 78% from the PCV7 to the PCV13 period, resulting in an overall 82% reduction (Table 2, Figure 1). Rates of M. catarrhalis and S. pyogenes did not change significantly during the PCV7 period compared with the pre- PCV period. However, during the PCV13 period, 81% and 62% reduction of the above two pathogens were observed, respectively. Culture-negative OM rates declined by 13% from the pre- PCV to the PCV7 period and by 44% from the PCV7 to the PCV13 period, resulting in an overall reduction of 51% throughout the study (Table 2, Figure 1). Table 1. Total Number of Otitis Media Episodes in Children Aged <4, 4 11, 12 23, 24 35, and <36 Months in Southern Israel, July 2004 Through June 2015 Age (mo) Study Period a All Pnc Mixed Pnc-NTHi Single NTHi Other Positive b Culture Negative All-Cause Otitis Media <4 Pre-PCV PCV PCV July 2004 June Pre-PCV PCV PCV July 2004 June Pre-PCV PCV PCV July 2004 June Pre-PCV PCV PCV July 2004 June All children <36 Pre-PCV PCV PCV July 2004 June Abbreviations: NTHi, nontypable Haemophilus influenzae; PCV, pneumococcal conjugated vaccine; Pnc, Streptococcus pneumoniae. a The study periods are as follows: pre-pcv, July 2004 June 2008; PCV7, July 2009 June 2011; and PCV13, July 2013 June b Moraxella catarrhalis, Streptococcus pyogenes, and mixed infections other than Pnc + NTHi infection. PCV and Nonpneumococcal Otitis Media CID 2016:63 (1 September) 613

4 Table 2. Impact of the Sequential Pneumococcal Conjugated Vaccine (PCV)7/PCV13 Introduction to the National Immunization Plan on Pneumococcal, Nonpneumococcal, and All-Cause Otitis Media Rates in Children Aged <36 Months Episode by Pathogen Pre-PCV a Mean Annual PCV7 a Mean Annual PCV7 vs Pre-PCV PCV13 vs PCV7 Pneumococcal VT7 + 6A 4.1 ± 0.3 (n = 718) 1.5 ± 0.4 (n = 134) 0.35 (.29.42) 0.1 ± 0.0 (n = 9) 0.06 (.03.12) 0.02 (.01.04) serotypes Pneumococcal VT5 1.5 ± 0.1 (n = 261) 1.6 ± 0.5 (n = 148) 1.07 ( ) 0.2 ± 0.0 (n = 19) 0.12 (.07.19) 0.12 (.08.20) serotypes Pneumococcal non- 1.1 ± 0.2 (n = 193) 1.6 ± 0.4 (n = 152) 1.48 ( ) 0.9 ± 0.3 (n = 92) 0.55 (.42.71) 0.82 ( ) VT13 serotypes All pneumococcal 6.8 ± 0.5 (n = 1178) 4.7 ± 0.5 (n = 436) 0.70 (.63.78) 1.2 ± 0.3 (n = 122) 0.25 (.21.31) 0.18 (.15.21) serotypes NTHi + Streptococcus 2.7 ± 0.3 (n = 461) 2.1 ± 0.2 (n = 196) 0.80 (.68.95) 0.5 ± 0.2 (n = 48) 0.22 (.16.31) 0.18 (.13.24) pneumoniae NTHi single culture 4.9 ± 0.5 (n = 853) 4.6 ± 0.5 (n = 422) 0.93 ( ) 1.5 ± 0.2 (n = 148) 0.32 (.26.38) 0.30 (.25.35) All NTHi 7.9 ± 0.7 (n = 1370) 7.2 ± 0.8 (n = 664) 0.91 ( ) 2.0 ± 0.4 (n = 203) 0.28 (.24.33) 0.25 (.22.29) Overall OM Overall OM in all children aged <3 years significantly declined from the pre-pcv rate of 19.6 ± 2.5 per 1000 by 15% during the PCV7 period and by 62% from PCV7 to PCV13 period, resulting in an overall 68% reduction throughout the study (Table 2, Figure 1). PCV Impact by Age Rates of OM caused by VT7 + 6A were reduced significantly in the 3 age groups (<12, 12 23, and months) during the PCV7 period, with a somewhat flatter decline in children aged months (Table 3). During the PCV13 period, VT A OM continued to decline substantially, reaching approximately 90% reduction in all ages. In the PCV7 period, OM caused by VT5 did not change significantly in any age group compared with the pre-pcv period. However, during the PCV13 period, VT5 rates declined substantially in all ages: 92%, 86%, and 80% in children aged <12, 12 23, and months, respectively, compared with the PCV7 period. During the PCV7 period, the overall pneumococcal OM rate declined significantly in children aged <24 months but not in children aged months. However, during the PCV13 period, the reductions were significant in all ages. In contrast to pneumococcal OM, nonpneumococcal OM trends were not as homogenous when analyzed by age. Rates of NTHi, culture-negative, and all-cause OM declined significantly, mainly in children aged <12 months during the PCV7 period. However, during the PCV13 period, significant reductions were observed for all ages. Since infants aged <4 months were expected to receive 1 PCV dose, rate reductions in this age group depended mainly on indirect protection and thus were expected to be observed somewhat later than for children aged 4 11 months. Indeed, during the PCV7 period, both pneumococcal and NTHi OM rates were significantly reduced in children aged 4 11 months but not in those aged <4 months. However, during the PCV13 period, which mostly spanned the period that was 3 years post-pcv introduction, overall reductions reached similar magnitude between the 2 groups. OM in Jewish vs Bedouin Children During the PCV7 period, pneumococcal and all-cause OM declined similarly and significantly in both ethnic groups (Table 4). In contrast, NTHi and culture-negative OM declined significantly during the PCV7 period in Bedouin children only. During the PCV13 period, all categories of OM declined in both Jewish and Bedouin children; still, the decline among Bedouin infants was more pronounced. DISCUSSION PCV13 a Mean Annual Incidence b (±SD) PCV13 vs Pre- PCV Single Moraxella 0.3 ± 0.1 (n = 54) 0.3 ± 0.1 (n = 30) 1.05 ( ) 0.1 ± 0.0 (n = 6) 0.18 (.08.44) 0.19 (.08.44) catarrhalis Single Streptococcus 0.6 ± 0.0 (n = 109) 0.7 ± 0.2 (n = 63) 1.09 ( ) 0.2 ± 0.0 (n = 24) 0.35 (.22.55) 0.38 (.24.59) pyogenes Culture negative 6.7 ± 1.6 (n = 1159) 5.8 ± 0.2 (n = 535) 0.87 (.79.96) 3.2 ± 0.2 (n = 329) 0.56 (.49.64) 0.49 (.43.55) All-cause otitis media 19.6 ± 2.5 (n = 3411) 16.6 ± 1.5 (n = 1532) 0.85 (.80.90) 6.3 ± 0.3 (n = 636) 0.38 (.34.41) 0.32 (.29.35) Each study year is 1 July through 30 June. Bold type represents numbers which are Statistically significant (P <.05). Abbreviations: CI, confidence interval; IRR, incidence rate ratio; NTHi, nontypable Haemophilus influenzae; PCV, pneumococcal conjugated vaccine; SD, standard deviation; VT, vaccine serotypes. a The study periods are as follows: pre-pcv, July 2004 June 2008; PCV7, July 2009 June 2011; and PCV13, July 2013 June b Incidence per 1000 children aged <36 months. Following the sequential introduction of PCV7/PCV13, rates of OM episodes enriched with complex cases submitted for MEF culture were substantially reduced in children aged <3 years in southern Israel. Within 6 years of introduction, episodes caused 614 CID 2016:63 (1 September) Ben-Shimol et al

5 Figure 1. Incidence dynamics of pneumococcal, nontypable Haemophilus influenzae, culture-negative, and all-cause otitis media episodes during which middle ear fluid culture was obtained in children aged <36 months in southern Israel, July 2004 through June Abbreviations: NTHi, nontypable Haemophilus influenzae; OM, otitis media; PCV, pneumococcal conjugated vaccine; Pnc, Streptococcus pneumoniae. by S. pneumoniae, NTHi,M. catarrhalis, S. pyogenes, mixed NTHi + S. pneumoniae infection, and culture-negative episodes were reduced by 82%, 75%, 81%, 62%, 82%, and 51%, respectively. An overall 68% reduction in all-cause episodes was observed. The reduction in pneumococcal OM following PCV introduction, mainly driven by near elimination of VT13 episodes, was expected. However, the remarkable reduction observed in nonpneumococcal OM episodes might seem paradoxical. We hypothesize that through prevention of early OM, mainly caused by PCV7/PCV13 pneumococci, subsequent complex OM episodes were prevented. This was associated with a reduction in additional pathogens, since complex OM cases are often polymicrobial. This is further supported by previous studies in which the efficacy of a H. influenzae-derived protein D conjugated to pneumococcal polysaccharides vaccine (PCV10) was greater (although not statistically significant) for recurrent OM (56%) and ventilation tube placement (60%) than for any OM episode (34%) [21]. Furthermore, previous studies have shown that culture-negative, complex OM frequently contains bacterial pathogens associated with biofilm formation and characterized by a paucity of MEF planktonic forms [7 10, 13]. Biofilms were also implicated in chronic bacterial infections such as rhino sinusitis [12, 22], cystic fibrosis [14], and recurrent/chronic lower respiratory infections [13 15]. Our hypothesis is strongly supported by the differential impact of vaccination on pneumococcal vs nonpneumococcal disease. As expected, the impact PCV on pneumococcal OM was greater and faster than that on OM caused by other pathogens and driven by the >90% reduction of VT13 diseases. Furthermore, PCV7 introduction was associated with the reduction of only VT7 + 6A, and the sequential PCV13 introduction resulted in reduction of the additional VT5. Consistently, non- VT13 pneumococcal OM cases were not reduced until the second study period (PCV13 period) when nonpneumococcal OM episodes were also markedly reduced. This was in accordance with other reports [1], emphasizing the primary role of PCVs in this reduction. PCV and Nonpneumococcal Otitis Media CID 2016:63 (1 September) 615

6 Table 3. Impact of the Sequential Pneumococcal Conjugated Vaccine (PCV)7/PCV13 Introduction to the National Immunization Plan on Pneumococcal, Nontypable Haemophilus Influenzae, Culture Negative, and All-Cause Otitis Media Rates in Children Aged <12, 12 23, and Months OM Caused by Age (mo) Pre-PCV a Mean Annual Incidence (±SD) PCV7 a Mean Annual Incidence (±SD) PCV7 vs Pre-PCV PCV13 vs PCV7 VT7 + 6A < ± 0.7 (n = 378) 2.3 ± 0.4 (n = 73) 0.36 (.28.46) 0.1 ± 0.0 (n = 4) 0.05 (.02.14) 0.02 (.01.05) <4 0.9 ± 0.3 (n = 52) 0.8 ± 0.5 (n = 25) 0.89 ( ) 0.0 ± 0.0 (n = 0) 0.02 (.00.29) 0.02 (.00.26) ± 0.4 (n = 326) 1.5 ± 0.1 (n = 48) 0.27 (.20.37) 0.1 ± 0.0 (n = 4) 0.08 (.03.21) 0.02 (.01.06) ± 0.4 (n = 283) 1.5 ± 0.3 (n = 45) 0.30 (.22.41) 0.1 ± 0.0 (n = 5) 0.10 (.04.25) 0.03 (.01.07) ± 0.3 (n = 57) 0.5 ± 0.3 (n = 16) 0.54 (.31.94) 0.0 ± 0.0 (n = 0) 0.03 (.00.46) 0.02 (.00.24) VT5 < ± 0.5 (n = 158) 2.7 ± 1.0 (n = 86) 1.00 ( ) 0.2 ± 0 (n = 8) 0.08 (.04.17) 0.08 (.04.17) <4 0.8 ± 0.2 (n = 47) 1.1 ± 0.6 (n = 36) 1.41 ( ) 0.1 ± 0 (n = 5) 0.13 (.05.32) 0.18 (.07.45) ± 0.3 (n = 111) 1.6 ± 0.4 (n = 50) 0.83 ( ) 0.1 ± 0 (n = 3) 0.05 (.02.17) 0.05 (.01.14) ± 0.3 (n = 84) 1.4 ± 0.5 (n = 44) 0.99 ( ) 0.2 ± 0.1 (n = 7) 0.14 (.06.32) 0.14 (.07.31) ± 0.1 (n = 19) 0.6 ± 0.0 (n = 18) 1.83 ( ) 0.1 ± 0 (n = 4) 0.20 (.07.59) 0.37 ( ) Streptococcus pneumoniae Further support for our hypothesis comes from the observed differential impact by age group. The decline in all pneumococcal OM and all-cause OM was observed earlier and in the greatest magnitude in children aged <12 months. Furthermore, even within this age group, the reduction was observed earlier in children aged 4 11 months who were more likely to have had 2 PCV doses compared with those aged <4 months who received 1 PCV dose. In the latter age group, protection depends almost entirely on indirect effect through reduction in carriage among older, vaccinated children, which is not maximized before a considerable proportion of the population is appropriately vaccinated [23]. Similarly, PCV impact was delayed in children aged months who (at least during the early post-pcv period) were unlikely to be immunized during early infancy; it is likely that a large proportion of these older children were immunized after the first episode of OM had already occurred (during a catch-up PCV13 a Mean Annual Incidence (±SD) PCV13 vs Pre- PCV < ± 1.3 (n = 661) 7.8 ± 1.0 (n = 245) 0.68 (.59.79) 1.4 ± 0.5 (n = 47) 0.17 (.13.24) 0.12 (.09.16) <4 2.7 ± 0.8 (n = 154) 2.7 ± 0.1 (n = 86) 1.03 ( ) 0.3 ± 0.1 (n = 9) 0.09 (.05.19) 0.10 (.05.19) ± 0.6 (n = 507) 5.1 ± 0.9 (n = 159) 0.58 (.48.69) 1.1 ± 0.4 (n = 38) 0.22 (.15.31) 0.13 (.09.17) ± 0.5 (n = 417) 4.4 ± 0.5 (n = 136) 0.61 (.51.75) 1.7 ± 0.4 (n = 59) 0.39 (.29.53) 0.24 (.18.32) ± 0.4 (n = 100) 1.8 ± 0.0 (n = 55) 1.06 ( ) 0.5 ± 0.1 (n = 16) 0.26 (.15.46) 0.28 (.16.47) Nontypable < ± 1.7 (n = 729) 10.3 ± 1.2 (n = 323) 0.82 (.72.93) 2.2 ± 0.6 (n = 77) 0.22 (.17.28) 0.18 (.14.22) Haemophilus <4 2.5 ± 0.9 (n = 144) 2.1 ± 0.7 (n = 65) 0.83 ( ) 0.2 ± 0.2 (n = 6) 0.08 (.04.19) 0.07 (.03.16) influenzae ± 1.1 (n = 585) 8.2 ± 0.4 (n = 258) 0.81 (.70.94) 2.1 ± 0.4 (n = 71) 0.25 (.19.32) 0.20 (.16.26) ± 0.4 (n = 537) 8.8 ± 1.1 (n = 271) 0.95 ( ) 3.2 ± 0.3 (n = 108) 0.36 (.29.45) 0.34 (.28.42) ± 0.7 (n = 104) 2.3 ± 0.0 (n = 70) 1.30 ( ) 0.5 ± 0.3 (n = 18) 0.23 (.14.39) 0.3 (.18.50) Culture-negative < ± 3.6 (n = 638) 8.5 ± 0.9 (n = 266) 0.77 (.67.89) 4.6 ± 0.4 (n = 159) 0.54 (.45.66) 0.42 (.35.50) <4 2.8 ± 1.2 (n = 164) 2.0 ± 0.7 (n = 63) 0.71 (.53.95) 1.1 ± 0.2 (n = 39) 0.56 (.38.84) 0.40 (.28.56) ± 2.4 (n = 474) 6.5 ± 0.1 (n = 203) 0.79 (.67.93) 3.5 ± 0.2 (n = 120) 0.54 (.43.67) 0.42 (.35.52) ± 1.1 (n = 389) 6.6 ± 0.9 (n = 203) 0.98 ( ) 3.6 ± 0.1 (n = 121) 0.54 (.43.68) 0.53 (.43.65) ± 0.6 (n = 132) 2.2 ± 0.5 (n = 66) 0.96 ( ) 1.5 ± 0.4 (n = 49) 0.67 (.46.97) 0.65 (.47.90) All-cause < ± 5.8 (n = 1880) 24.5 ± 2.4 (n = 769) 0.75 (.69.82) 8.1 ± 0.4 (n = 280) 0.33 (.29.38) 0.25 (.22.28) <4 7.6 ± 2.5 (n = 438) 6.5 ± 1.4 (n = 203) 0.86 ( ) 1.6 ± 0 (n = 54) 0.24 (.18.33) 0.21 (.16.27) ± 3.5 (n = 1442) 18.0 ± 1.0 (n = 566) 0.72 (.66.80) 6.5 ± 0.4 (n = 226) 0.36 (.31.42) 0.26 (.23.30) ± 1.5 (n = 1225) 18.9 ± 1.0 (n = 582) 0.90 (.81.99) 8.2 ± 0.5 (n = 277) 0.43 ( ) 0.39 (.34.44) ± 1.2 (n = 306) 6.0 ± 0.9 (n = 181) 1.14 ( ) 2.4 ± 0.1 (n = 79) 0.39 (.30.51) 0.45 (.35.57) Bold type represents numbers which are Statistically significant (P <.05). Abbreviations: CI, confidence interval; IRR, incidence rate ratio; OM, otitis media; PCV, pneumococcal conjugated vaccine; SD, standard deviation; VT, vaccine serotypes. a The study periods are as follows: pre-pcv, July 2004 June 2008; PCV7, July 2009 June 2011; and PCV13, July 2013 June campaign). As shown previously, once recurrent OM is established, PCVs may not reduce or may even increase OM episodes, largely of non-pcv13 serotypes [24, 25]. A recent article from Finland supports these findings. In that study, protection against ventilation tube insertion tended to depend on age at vaccination: 13%, 11%, and 1% reduction of those vaccinated at age 6, 7 11, and months, respectively [26]. However, it is possible that the relatively minor effect on the youngest and oldest age groups (<4 months and months) could be due to the overall lower incidence of OM in these age groups. Vaccine impact was similar in the Jewish and Bedouin populations, although the impact was somewhat greater among Bedouin children, mainly during the late PCV13 period. This finding is important since in the crowded Bedouin population (in this aspect, resembling a developing-world population), PCV impact could, in theory, be delayed due to a higher 616 CID 2016:63 (1 September) Ben-Shimol et al

7 Table 4. Impact of the Sequential Pneumococcal Conjugated Vaccine (PCV)7/PCV13 Introduction to the National Immunization Plan on All-Cause, Pneumococcal, Nontypable Haemophilus influenzae, Other Bacterial, and Culture-Negative Otitis Media Rates in Jewish and Bedouin Children Aged <36 Months OM Type All pneumococcal OM All pneumococcal OM All NTHi OM All NTHi OM Other OM c Other OM c Culture-negative OM Culture-negative OM All-cause OM All-cause OM Pre-PCV a Mean Annual PCV7 a Mean Annual carriage rate, starting from a very young age and expanding to older age, than in the Jewish population, resulting in higher exposure rates both to NTHi and S. pneumoniae [27 32]. We previously showed that PCV7/PCV13 reduced vaccine-type (VT) carriage in the 2 ethnic groups in a similar fashion, despite large differences in carriage rates [30, 33, 34]. Our results are therefore reassuring and suggest that in developing populations where chronic OM cases (including chronic suppurative OM and hearing loss) are prevalent [35], PCVs may prevent a sizeable proportion of the sequelae. Concerns were raised regarding potential post-pcv compensatory increase in pneumococcal disease caused by non-vt serotypes (replacement disease) [33, 36, 37]. However, in previous studies that dealt with OM likely to be enriched with complex OM (the POETand COMPAS studies), no significant increase in non-vt disease was observed in vaccinated children [21]. Our group and others have shown an increase in non-vt pneumococcal nasopharyngeal carriage after PCV introduction [34, 38]. In contrast, no such increase was shown in complex OM. This is in accordance with a recent study from the United States in which pneumococcal isolates from middle ear or mastoid cultures from children were evaluated and no significant increase in non-vt was observed [39]. We speculate that a reduced host susceptibility to non-vt (generally less invasive) disease derived from prevention of primary damage caused by the generally more invasive VT PCV7 vs Pre-PCV PCV13 a Mean Annual PCV13 vs PCV7 PCV13 vs Pre-PCV 7.9 ± 0.5 (n = 659) 5.0 ± 1.2 (n = 223) 0.64 (.55.74) 1.8 ± 0.5 (n = 86) 0.35 (.27.45) 0.22 (.18.28) 5.7 ± 0.6 (n = 519) 4.5 ± 0.1 (n = 213) 0.78 (.66.91) 0.7 ± 0.2 (n = 36) 0.16 (.11.22) 0.12 (.09.17) 8.6 ± 0.8 (n = 718) 8.5 ± 1 (n = 379) 0.99 ( ) 3.1 ± 0.8 (n = 150) 0.36 (.30.43) 0.35 (.30.42) 7.2 ± 0.9 (n = 652) 6.0 ± 0.5 (n = 285) 0.83 (.72.95) 1.0 ± 0.0 (n = 53) 0.17 (.13.23) 0.14 (.11.19) 1.5 ± 0.2 (n = 126) 1.4 ± 0.6 (n = 63) 0.94 ( ) 0.5 ± 0.0 (n = 24) 0.34 (.21.55) 0.32 (.21.50) 0.4 ± 0.0 (n = 37) 0.6 ± 0.0 (n = 30) 1.53 ( ) 0.1 ± 0.0 (n = 6) 0.18 (.08.44) 0.28 (.12.67) 6.5 ± 1.0 (n = 541) 6.8 ± 0.1 (n = 304) 1.06 ( ) 4.0 ± 0.2 (n = 199) 0.59 (.49.71) 0.62 (.53.73) 6.8 ± 2.6 (n = 618) 4.8 ± 0.3 (n = 231) 0.71 (.61.82) 2.5 ± 0.6 (n = 130) 0.52 (.42.65) 0.37 (.30.44) 21.8 ± 2.3 (n = 1824) 19.9 ± 2.2 (n = 884) 0.91 (.84.99) 8.7 ± 1.1 (n = 427) 0.44 (.39.49) 0.40 (.36.44) 17.5 ± 3.7 (n = 1587) 13.6 ± 0.8 (n = 648) 0.77 (.71.85) 4.0 ± 0.5 (n = 209) 0.3 (.26.35) 0.23 (.20.27) Each study year is 1 July through 30 June. Bold type represents numbers which are Statistically significant (P <.05). Abbreviations: CI, confidence interval; IRR, incidence rate ratio; NTHi, nontypable Haemophilus influenzae; OM, otitis media; PCV, pneumococcal conjugated vaccine; SD, standard deviation. a The study periods are as follows: pre-pcv, July 2004 June 2008; PCV7, July 2009 June 2011; and PCV13, July 2013 June b Incidence per 1000 children. c Other OM: Moraxella catarrhalis and Streptococcus pyogenes. serotypes [16, 38] is responsible for the absence of non-vt OM increase, despite a parallel increase in nasopharyngeal carriage. The major strengths of our study are the prospective, active, and population-based conduct and the large number of cases. The main limitation of our study lies in the inability to estimate true OM rates, as our patients were enrolled only if MEF cultures were obtained. In that regard, it is possible that although indications for referral, tympanocentesis, or MEF culture did not change during the study period in our hospital, the perception of some physicians that OM rates declined following PCV introduction may have led to lower referral rates. However, the respective reduction of PCV7 serotypes followed by the additional PCV13 serotypes and the nonsignificant reduction of non-pcv13 serotypes oppose this possibility. Furthermore, the bacteriology of these children may not represent that of children with less severe or noncomplex disease. However, our complex OM-enriched population probably carries the main burden in OM [17]. In summary, this prospective, population-based, long-term, active surveillance demonstrates a rapid substantial decline in the incidence of both pneumococcal and nonpneumococcal OM episodes for which MEF cultures were obtained in children <3 years, following sequential PCV7/PCV13 introduction. The reduction in nonpneumococcal episodes may be attributed to the prevention of early OM episodes caused by the relatively PCV and Nonpneumococcal Otitis Media CID 2016:63 (1 September) 617

8 invasive PCV13-serotypes, resulting in a lower rate of complex, often nonpneumococcal OM. Notes Financial support. The study was funded in part by a grant from Pfizer (grant number 0887X1-4603). Potential conflicts of interest. S. B.-S. has received speakers fees from Pfizer. D. G. has received grants from Merck Sharp & Dohme; has been a scientific consultant for GlaxoSmithKline (GSK), Merck Sharp & Dohme, and Pfizer; and is a speaker for GSK, Merck Sharp & Dohme, and Pfizer. R. D. has received grants and research support from Pfizer and Merck Sharp & Dohme; has been a scientific consultant for Genocea, MeMed, Merck Sharp & Dohme, and Pfizer; and is a speaker for GSK and Pfizer. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. Dagan R, Pelton S, Bakaletz L, Cohen R. Prevention of early episodes of otitis media by pneumococcal vaccines might reduce progression to complex disease. Lancet Infect Dis 2016; 16: Sommerfleck P, Gonzalez Macchi ME, Pellegrini S, et al. Acute otitis media in infants younger than three months not vaccinated against Streptococcus pneumoniae. Int J Pediatr Otorhinolaryngol 2013; 77: Leibovitz E, Serebro M, Givon-Lavi N, et al. Epidemiologic and microbiologic characteristics of culture-positive spontaneous otorrhea in children with acute otitis media. Pediatr Infect Dis J 2009; 28: Casey JR, Adlowitz DG, Pichichero ME. New patterns in the otopathogens causing acute otitis media six to eight years after introduction of pneumococcal conjugate vaccine. Pediatr Infect Dis J 2010; 29: Naranjo L, Suarez JA, DeAntonio R, et al. Non-capsulated and capsulated Haemophilus influenzae in children with acute otitis media in Venezuela: a prospective epidemiological study. BMC Infect Dis 2012; 12: Wiertsema SP, Kirkham LA, Corscadden KJ, et al. Predominance of nontypeable Haemophilus influenzae in children with otitis media following introduction of a 3+0 pneumococcal conjugate vaccine schedule. Vaccine 2011; 29: Barkai G, Leibovitz E, Givon-Lavi N, Dagan R. Potential contribution by nontypable Haemophilus influenzae in protracted and recurrent acute otitis media. Pediatr Infect Dis J 2009; 28: Moriyama S, Hotomi M, Shimada J, Billal DS, Fujihara K, Yamanaka N. Formation of biofilm by Haemophilus influenzae isolated from pediatric intractable otitis media. Auris Nasus Larynx 2009; 36: Hall-Stoodley L, Hu FZ, Gieseke A, et al. Direct detection of bacterial biofilms on the middle-ear mucosa of children with chronic otitis media. JAMA 2006; 296: Hotomi M, Arai J, Billal DS, et al. Nontypeable Haemophilus influenzae isolated from intractable acute otitis media internalized into cultured human epithelial cells. Auris Nasus Larynx 2010; 37: Murphy TF, Faden H, Bakaletz LO, et al. Nontypeable Haemophilus influenzae as a pathogen in children. Pediatr Infect Dis J 2009; 28: Galli J, Calo L, Ardito F, et al. Damage to ciliated epithelium in chronic rhinosinusitis: what is the role of bacterial biofilms? Ann Otol Rhinol Laryngol 2008; 117: Eldika N, Sethi S. Role of nontypeable Haemophilus influenzae in exacerbations and progression of chronic obstructive pulmonary disease. Curr Opin Pulm Med 2006; 12: Craven V, Everard ML. Protracted bacterial bronchitis: reinventing an old disease. Arch Dis Child 2013; 98: Zhang L, Xie J, Patel M, et al. Nontypeable Haemophilus influenzae genetic islands associated with chronic pulmonary infection. PLoS One 2012; 7:e Dagan R, Leibovitz E, Greenberg D, Bakaletz L, Givon-Lavi N. Mixed pneumococcal-nontypeable Haemophilus influenzae otitis media is a distinct clinical entity with unique epidemiologic characteristics and pneumococcal serotype distribution. J Infect Dis 2013; 208: Ben-Shimol S, Givon-Lavi N, Leibovitz E, Raiz S, Greenberg D, Dagan R. Nearelimination of otitis media caused by 13-valent pneumococcal conjugate vaccine (PCV) serotypes in southern Israel shortly after sequential introduction of 7-valent/13-valent PCV. Clin Infect Dis 2014; 59: Dagan R, Leibovitz E, Fliss DM, et al. Bacteriologic efficacies of oral azithromycin and oral cefaclor in treatment of acute otitis media in infants and young children. Antimicrob Agents Chemother 2000; 44: Ben-Shimol S, Greenberg D, Givon-Lavi N, et al. Rapid reduction in invasive pneumococcal disease after introduction of PCV7 into the National Immunization Plan in Israel. Vaccine 2012; 30: Israel Central Bureau of Statistics. Statistical Abstract of Israel No. 63, Prymula R, Peeters P, Chrobok V, et al. Pneumococcal capsular polysaccharides conjugated to protein D for prevention of acute otitis media caused by both Streptococcus pneumoniae and non-typable Haemophilus influenzae: a randomised double-blind efficacy study. Lancet 2006; 367: Chan J, Hadley J. The microbiology of chronic rhinosinusitis: results of a community surveillance study. Ear Nose Throat J 2001; 80: Scarbrough Lefebvre CD, Terlinden A, Standaert B. Dissecting the indirect effects caused by vaccines into the basic elements. Hum Vaccin Immunother 2015; 11: Veenhoven R, Bogaert D, Uiterwaal C, et al. Effect of conjugate pneumococcal vaccine followed by polysaccharide pneumococcal vaccine on recurrent acute otitis media: a randomized study. Lancet 2003; 361: van Kempen MJ, Vermeiren JS, Vaneechoutte M, et al. Pneumococcal conjugate vaccination in children with recurrent acute otitis media: a therapeutic alternative? Int J Pediatr Otorhinolaryngol 2006; 70: Palmu AA, Jokinen J, Nieminen H, et al. Effectiveness of the ten-valent pneumococcal conjugate vaccine against tympanostomy tube placements in a cluster-randomized trial. Pediatr Infect Dis J 2015; 34: Woolfson A, Huebner R, Wasas A, Chola S, Godfrey-Faussett P, Klugman K. Nasopharyngeal carriage of community-acquired, antibiotic-resistant Streptococcus pneumoniae in a Zambian paediatric population. Bull World Health Organ 1997; 75: Denno DM, Frimpong E, Gregory M, Steele RW. Nasopharyngeal carriage and susceptibility patterns of Streptococcus pneumoniae in Kumasi, Ghana. West Afr J Med 2002; 21: Erasto P, Hoti F, Granat SM, Mia Z, Makela PH, Auranen K. Modelling multi-type transmission of pneumococcal carriage in Bangladeshi families. Epidemiol Infect 2010; 138: Greenberg D, Givon-Lavi N, Ben Shimol S, Dagan R. Primary school children constitute an important pneumococcal vaccine serotype (VT) reservoir, 5 years after initiation of widespread pneumococcal conjugate vaccine (PCV) program. In: IDweek October 2015, Philadelphia, PA. 31. Reis JN, Palma T, Ribeiro GS, et al. Transmission of Streptococcus pneumoniae in an urban slum community. J Infect 2008; 57: Lewnard J, Givon-Lavi N, Huppert A, et al. Epidemiological markers for interactions among Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus in upper respiratory tract carriage. J Infect Dis 2016; 213: Weil-Olivier C, van der Linden M, de Schutter I, Dagan R, Mantovani L. Prevention of pneumococcal diseases in the post-seven valent vaccine era: a European perspective. BMC Infect Dis 2012; 12: Ben Shimol S, Givon-Lavi N, Greenberg D, Dagan R. Pneumococcal nasopharyngeal carriage in children <5 years of age visiting the pediatric emergency room in relation to PCV7 and PCV13 introduction in southern Israel. Hum Vaccin Immunother 2016; 12: Vergison A, Dagan R, Arguedas A, et al. Otitis media and its consequences: beyond the earache. Lancet Infect Dis 2010; 10: Eskola J, Kilpi T, Palmu A, et al. Efficacy of a pneumococcal conjugate vaccine against acute otitis media. N Engl J Med 2001; 344: Block SL, Hedrick J, Harrison CJ, et al. Community-wide vaccination with the heptavalent pneumococcal conjugate significantly alters the microbiology of acute otitis media. Pediatr Infect Dis J 2004; 23: Flasche S, Van Hoek A, Sheasby E, et al. Effect of pneumococcal conjugate vaccination on serotype-specific carriage and invasive disease in England: a cross-sectional study. PLoS Med 2011; 8:e Kaplan SL, Center KJ, Barson WJ, et al. Multicenter surveillance of Streptococcus pneumoniae isolates from middle ear and mastoid cultures in the 13-valent pneumococcal conjugate vaccine era. Clin Infect Dis 2015; 60: CID 2016:63 (1 September) Ben-Shimol et al