Streptococcus pneumoniae is a leading cause of pediatric bacteremia and meningitis, with the highest incidence occurring in children < 5 years old.1,2 Occult bacteremia is defined as the presence of bacteria (90% are caused by S. pneumoniae) in the bloodstream of a febrile child (usually 3–36 months old) who looks well (nontoxic/nonseptic appearance) and has no obvious source of infection.3,4 Most children with occult pneumococcal bacteremia improve spontaneously, but ~25% of untreated patients have persistent bacteremia or develop new focal infections, including 3–6% who develop meningitis.3,4 Thus, in the clinical settings, children with suspected occult bacteremia are empirically treated with antibiotics to prevent those complications.3,4
The implementation of the 7-valent and 13-valent conjugate pneumococcal vaccines (PCV7 and PCV13) in National Immunization Programs (NIPs) resulted in a rapid and substantial reduction of invasive pneumococcal disease (IPD) rates in children.5–9 However, IPD is not a single clinical syndrome, and thus, PCV impact may not be similar in all outcomes within IPD.10–12
In Israel, of all IPD episodes in children < 5 years old in the pre-PCV era, 48.5% were diagnosed as sepsis/bacteremia (including septic/severe cases but also occult bacteremia cases).10 Other diagnoses included Bacteremic pneumonia (35.5%), meningitis (9.4%), bacteremic cellulitis, septic arthritis, peritonitis and sinusitis (2.9% grouped) and 3.7% had a missing diagnosis. Importantly, even in the Israeli setting, where all IPD episodes in children are monitored nationwide for > 25 years, it is hard to establish the diagnosis of occult bacteremia, as often clinical data regarding toxic/septic appearance and focal findings are missing.
Several studies reported reduction in occult bacteremia following PCVs introduction.13–17 However, these studies often had limitations in assessment of the real occult bacteremia incidence rate, as data regarding clinical presentation were usually not prospectively obtained, and consequently the prevalence (proportions) of pathogen-specific bacteremia out of all blood cultures obtained were reported. We previously evaluated PCV impact on incidence rates of different IPD end points10–12 and found that following PCV7/PCV13 introduction, dynamic rates of pneumococcal meningitis, bacteremic pneumonia and other (nonmeningitis, nonpneumonia) IPD episodes differed because of age and serotype distribution variability among the various IPD groups.
Occult bacteremia is essentially an IPD without a focus in a nontoxic child, mostly with a favorable outcome. To better understand PCVs impact on various IPD end points, we divided all IPD episodes into 2 groups: severe IPD (SIPD, including episodes with prolonged hospitalization, as well as cases with meningitis, pneumonia and mastoiditis as a focus) and nonsevere IPD (NSIPD, including episodes with a favorable outcome, without recognized meningitis, pneumonia and mastoiditis as a focus). These definitions were made to identify children with occult pneumococcal bacteremia and other nonsevere presentations of pneumococcal infections.
We assessed the dynamics of NSIPD and SIPD rates in children < 5 years in Israel before and after PCV7/PCV13 implementation.
Setting and Study Population
The study population comprised all children < 5 years old in Israel. As of 2004, 2009 and 2014, Israel had a population of ~699,000, 754,000 and 850,000 children < 5-years-old, respectively.18
Local investigators in each center responded to a monthly distributed questionnaire sent by the principal investigator at the study headquarters located at the Pediatric Infectious Disease Unit of the Soroka University Medical Center.
An invasive pneumococcal disease episode was defined as an illness during which S. pneumoniae was isolated from blood, CSF or both. Nonculture diagnoses (polymerase chain reaction, antigen testing, Gram stain results or clinical diagnosis) were excluded. In Israel, < 5% of all IPD episodes are diagnosed by nonculture methods. Positive cultures from sterile sites other than blood or CSF (i.e., joint fluid, pleural fluid) were also excluded.
To identify children with occult pneumococcal bacteremia and other nonsevere presentations of pneumococcal infections, 2 criteria were used: (1) Lack of identification of meningitis, pneumonia or mastoiditis as IPD foci; and (2) outcome parameters suggesting nonsevere disease. Thus, NSIPD episodes were defined as IPD episodes without meningitis, pneumonia or mastoiditis in a child with a favorable outcome. Favorable outcome was defined as an episode of a child who was discharged from the emergency room or admitted for < 5 days and was neither hospitalized in an intensive care unit nor died. Nonsevere invasive pneumococcal disease episodes mostly (~90%) comprised pneumococcal bacteremia without focus, while other (focal) NSIPD episodes included cellulitis, sinusitis and arthritis.10
All IPD cases not included in the NSIPD group were defined as SIPD. Therefore, SIPD cases included all meningitis, pneumonia or mastoiditis episodes, as well as bacteremia episodes with prolonged hospitalization, intensive care unit hospitalization and/or mortality.
PCV7/PCV13 Vaccine Uptake
The introduction of the 7-valent pneumococcal conjugate vaccine into the NIP was initiated in July 2009 with a catch-up campaign in children < 2 years.10,11 In November 2010, PCV13 replaced PCV7, without further catch-up.
By June 2011 and December 2012, 36% and 87%, respectively, of children 24–35 months, received ≥ 3 PCV7/PCV13 doses and > 90% received ≥ 3 PCV13 doses by June 2014 and June 2015.
Annual (July through June) incidence rates were calculated as the number of blood/CSF-positive culture cases divided by the total population at risk during each year of the study. The age-specific population at risk was estimated according to the Israeli Central Bureau of Statistics reports for the appropriate years.
For episodes in which serotype and/or serogroup were missing, a detailed extrapolation was conducted according to the distribution of serotypes in the age and ethnic groups during the same year, as described elsewhere.5
The proportion of isolates with serotype determination increased from ~60% in the pre-PCV period (2000–2008) to > 95% since 2009–2010.
To assess changes in NSIPD and SIPD incidence, we used annual rates. Three subperiods were defined, to allow better impact appreciation; (1) pre-PCV: July 1999–June 2008; (2) PCV7: July 2010–June 2011; and (3) PCV13: July 2013–June 2015. Mean incidences during PCV7 and PCV13 periods were compared with pre-PCV period.
Four age groups were defined; < 4, 4–11, 12–23 and 24–59 months. The subdivision of children < 12 months old group to < 4 and 4–11 months was done since children < 4 months are expected to have ≤ 1 PCV dose and thus can be counted as poorly or not vaccinated and are almost entirely dependent on indirect protection.
Incidences were calculated using the birth cohorts born in Israel, according to the Israeli Central Bureau of Statistics reports.18 Incidence rate ratios and 95% confidence intervals were calculated for NSIPD and SIPD and also analyzed by age groups.
Data were analyzed with SPSS 18.00 software. Univariate analyses were conducted using 2-tailed χ2 test or Student’s t test, where appropriate. P value < 0.05 was considered statistically significant.
During the study period, from July 1999 through June 2015, 4,457 IPD episodes in children < 5 years old were identified. Of these, 3,398 (76.2%) were defined as severe episodes, 1,022 (22.9%) as nonsevere episodes and in 37 (0.8%) episodes severity could not be ascertained (Table, Supplemental Digital Content 1, http://links.lww.com/INF/D183). Among the overall study population, 59% were males; 78% and 22% were Jewish and non-Jewish children, respectively. Children < 4, 4–11, 12–23 and 24–59 months old constituted 11%, 24%, 38% and 27% of the study population, respectively.
Severe IPD episodes included pneumonia (49%), meningitis (14%), sepsis (35%) and mastoiditis (< 2%). The case fatality rate was < 2%. In 90% of NSIPD episodes, no focus was identified (foci identified included cellulitis, arthritis and sinusitis).
Overall, 35% of all isolates were extrapolated for missing serotypes during the study period. The proportion of isolates with serotype determination increased from 61% in the pre-PCV period (1999–2008) to > 95% in the PCV era in 2009 through 2015. Thus, data regarding overall, severe and nonsevere IPD rates are presented for the entire study period (1999 through 2015), but data regarding specific serotypes/serotype groups are only presented for the period between 2009 and 2015.
Pre-PCV Period (July 1999 through June 2008) Characteristics
Children < 4, 4–11, 12–23 and 24–59 months old comprised 10.3%, 24.8%, 38.9% and 26.0% of all IPD, respectively.
The proportion of severe episodes in children < 4 months (84.9%) was significantly higher compared with children 4–11 months (72.3%; P < 0.001), 12–23 months (75.1%; P < 0.001) and 24–59 months (79.0%; P = 0.03).
Overall Nonsevere and Severe IPD Rate Dynamics (July 1999 through June 2015)
In the PCV7 period, compared with the pre-PCV period, NSIPD rates significantly declined by 52% (IRR = 0.48; 0.35–0.65), while SIPD rates declined less prominently by 24% (IRR = 0.76; 0.66–0.88; Table 1, Figs. 1, 2). In contrast, following PCV13 introduction, compared with the PCV7 period, NSIPD rates declined nonsignificantly by 17% (IRR = 0.83 [0.57–1.22]), while SIPD rates declined significantly further by additional 53% (IRR = 0.47; 0.39–0.57).
These trends resulted in an overall reduction (comparing the PCV13 and the pre-PCV periods) in NSIPD and SIPD rates by 60% (IRR = 0.40; 0.32–0.51) and 64% (IRR = 0.36; 0.32–0.42), respectively.
NSIPD and SIPD Rate Dynamics by Vaccine and Nonvaccine Serotypes (July 2009 through June 2015)
During the PCV13 period (2013–2015), for PCV7 serotypes and the 6 additional PCV13 serotypes, both SIPD and NSIPD rates substantially declined compared with the early-PCV period (2009–2010; Table 2). For PCV7 serotypes (7VT), NSIPD declined by 76% and for SIPD rates declined by 95%. For the 6 additional PCV13 serotypes (6VT), NSIPD declined by 92% and SIPD rates declined by 88% (IRR = 0.12 [0.07–0.18]). For non-PCV13 serotypes (non-VT13), both NSIPD and SIPD rates substantially increased during the PCV13 period; NSIPD rates increased by 74% and SIPD rates increased by 150%.
Proportions of 7VT, 6VT and Non-VT13 in SIPD and NSIPD in Children < 60 Months in Israel, 2009–2015
In the early-PCV period (2009–2010), in both NSIPD and SIPD, proportions of IPD caused by PCV7 serotypes and the 6 additional PCV13 serotypes were similar (Fig. 3). In the PCV7 period, proportions of IPD caused by 7VT were similar in NSIPD and SIPD. In contrast, proportions of IPD caused by additional-6VT were lower in NSIPD than in SIPD (51.3% and 70.4%, respectively, P = 0.025)
The differences in the proportions of 7VT, additional-6VT and non-VT13 observed in the PCV7 period disappeared in the PCV13 period.
Comparison of Rate Dynamics of NSIPD and SIPD Between Different Age Groups (July 1999 through June 2015)
Both NSIPD and SIPD rates declined significantly, comparing the pre-PCV period to the PCV13 period in all 4 age groups (Table, Supplemental Digital Content 2, http://links.lww.com/INF/D184). NSIPD and SIPD rates declined significantly also comparing the pre-PCV and PCV7 period for infants 4–11 and children 12–23 months of age, but not for infants < 4 months old and children 24–59 months old. Whereas significant reductions in SIPD rates comparing the PCV7 and the PCV13 periods were evident for children < 4, 12–23 and 24–59 months old, it was observed for NSIPD only in children 24–59 months old.
In the current study, both SIPD and NSIPD rates significantly declined (by 64% and 60%, respectively) following PCV7/PCV13 sequential introduction. However, the dynamics of those reductions were different. Initially, following PCV7 introduction, the decline observed in SIPD rates was less prominent than that observed in NSIPD rates. Subsequently, following PCV13 introduction, both SIPD and NSIPD rates declined resulting in an overall similar impact in the 2 groups.
The discrepancy between SIPD and NSIPD rate dynamics was mainly driven by the differences in serotype distribution between SIPD and NSIPD. In the early PCV era, NSIPD episodes were “enriched” with VT7 serotypes. Similarly, in SIPD, higher proportion of IPD caused by additional PCV13 serotypes (VT6) was observed in the early PCV era (compared with NSIPD). Thus, it is not surprising that NSIPD rates declined substantially immediately following PCV7 introduction, while the decline in SIPD rates was more substantial following PCV13 introduction.
Age is an important factor, beyond serotype distribution, on PCVs observed impact. In early PCV years, unvaccinated age groups (< 4 and 24–59 months old) were less impacted by PCVs, as they did not benefit from direct vaccine impact. These trends changed with time, and in the PCV13 period, all age groups were similarly impacted, indicating both direct and indirect (herd) protection. Additionally, in the < 4 month age group SIPD proportion of overall IPD is higher than in other age groups (due to higher proportion of meningitis and prolonged hospitalization durations), and this also contributed to the delayed PCV impact on the < 4-month-old age group.
In the current study, we assessed a different clinical syndrome, NSIPD (compared with SIPD), which we believe allows a better appreciation of PCVs impact on IPD. NSIPD, similarly to occult bacteremia, represents a group of IPD cases with favorable outcome. However, including other IPD episodes where a focus was identified (e.g., cellulitis) and also looking into outcome (e.g., hospitalization duration) allows a more complete understanding of the overall influence of PCVs on NSIPD.
It is important to be aware that the impressive decline in IPD rates (both severe and nonsevere), resulting from reduction in disease caused by vaccine serotypes (VT13), is accompanied by an increase in non-VT13 IPD rates. This phenomenon observed universally following PCVs introduction worldwide, derives from the replacement of vaccine serotypes carried in the nasopharynx by nonvaccine serotypes pneumococci.10,21–23 It is now well recognized that non-PCV13 serotypes are mostly of lower invasiveness compared with most PCV13 serotypes.21–23 However, as these nonvaccine serotypes become the predominant serotypes causing IPD, a continuous monitoring of their rates is mandated. Indeed, increased rates of IPD caused by serotype 12F and other nonvaccine serotypes have already been reported in Israel,24 as well as in the United Kingdom25 and Italy.26
Previous studies suggested, at least to some extent, physicians’ avoidance of prescribing antimicrobials in treatment protocols for febrile children 3–36 months old with leukocytosis, due to the reduction observed in occult bacteremia rates following PCVs introduction.13,19,27 We believe that physicians must be cautious when treating children with suspected occult bacteremia, and bear in mind that while disease rates are declining, nonvaccine serotype disease is increasing albeit not significantly. Furthermore, rates of meningitis, a possible complication of occult bacteremia, did not decline to the same extent as other IPD in Israel following PCV7/PCV13 sequential introduction, since most cases are caused by non-VT13 pneumococci.10 In consequence, a continuous surveillance is warranted to allow judicious modification, if any, to treatment protocols.
The main limitation of the current study is the relatively high rate of undetermined serotypes early in the pre-PCV period (missing ~40% of serotypes before 2009). However, in recent years, the rate of extrapolated serotypes has dropped to < 5% of all isolated pneumococci. Nevertheless, we acknowledge the possibility of secular trends and temporal fluctuations in pneumococcal serotypes in IPD, which may affect observed trends of SIPD and NSIPD, mainly in the pre-PCV era.
Additionally, some clinical data are missing not allowing the determination of “occult bacteremia” diagnosis. Nevertheless, we strongly believe that NSIPD definition is a more complete definition, which allows real appreciation of PCV impact on IPD episodes with favorable outcome.
We recognize the fact that SIPD and NSIPD are not always easy to define in the “real life setting.” It is conceivable that there is a “sliding scale” from NSIPD to severe and potentially fatal IPD, without a clear “cut-off” mark between the 2. Time is an important factor in this process, as a child with pneumococcal meningitis and pneumococcal bacteremia probably had “only” bacteremia a few hours earlier.
The main strengths of our study are the presence of long-term, nationwide pre-PCV IPD incidence data. Additionally, the relatively long time period that has elapsed since PCV13 introduction to the Israeli NIP (5 years) enables a genuine evaluation of PCV13 impact, as well as specific non-VT13 replacement rates.
In conclusion, following PCV7/PCV13 introduction, severe and nonsevere IPD rates were both substantially reduced. Different rate dynamics of the 2 groups were driven mainly by different serotype distribution, with PCV13 additional serotypes (additional-6VT) contributing in larger proportion to SIPD. These findings should be considered, and future surveillance studies are warranted when modification in treatment protocols for suspected occult bacteremia/IPD cases is contemplated.
1. Murray J, Agócs M, Serhan F, et alCenters for Disease Control and Prevention (CDC). Global invasive bacterial vaccine-preventable diseases surveillance—2008-2014. MMWR Morb Mortal Wkly Rep. 2014;63:1159–1162.
2. O’Brien KL, Wolfson LJ, Watt JP, et alHib and Pneumococcal Global Burden of Disease Study Team. Burden of disease caused by Streptococcus pneumoniae in children
younger than 5 years: global estimates. Lancet. 2009;374:893–902.
3. Kuppermann NOccult bacteremia in young febrile children
. Pediatr Clin North Am. 1999;46:1073–1109.
4. Rothrock SG, Harper MB, Green SM, et alDo oral antibiotics prevent meningitis and serious bacterial infections in children
with Streptococcus pneumoniae occult bacteremia
? A meta-analysis. Pediatrics. 1997;99:438–444.
5. Ben-Shimol S, Greenberg D, Givon-Lavi N, et alIsraeli Bacteremia and Meningitis Active Surveillance Group. Rapid reduction in invasive pneumococcal disease after introduction of PCV7 into the National Immunization Plan in Israel. Vaccine. 2012;30:6600–6607.
6. Ben-Shimol S, Greenberg D, Givon-Lavi N, et alEarly impact of sequential introduction of 7-valent and 13-valent pneumococcal conjugate vaccine on IPD in Israeli children
<5 years: an active prospective nationwide surveillance. Vaccine. 2014;32:3452–3459.
7. Carstairs KL, Tanen DA, Johnson AS, et alPneumococcal bacteremia in febrile infants presenting to the emergency department before and after the introduction of the heptavalent pneumococcal vaccine. Ann Emerg Med. 2007;49:772–777.
8. Herz AM, Greenhow TL, Alcantara J, et alChanging epidemiology of outpatient bacteremia in 3- to 36-month-old children
after the introduction of the heptavalent-conjugated pneumococcal vaccine. Pediatr Infect Dis J. 2006;25:293–300.
9. Tan TQPediatric invasive pneumococcal disease in the United States in the era of pneumococcal conjugate vaccines. Clin Microbiol Rev. 2012;25:409–419.
10. Ben-Shimol S, Givon-Lavi N, Grisaru-Soen G, et alComparative incidence dynamics and serotypes of meningitis, bacteremic pneumonia and other-IPD in young children
in the PCV era: insights from Israeli surveillance studies. Vaccine. 2018;36:5477–5484.
11. Ben-Shimol S, Greenberg D, Givon-Lavi N, et alIsrael Bacteremia and Meningitis Active Surveillance Group. Impact of PCV7/PCV13 introduction on invasive pneumococcal disease (IPD) in young children
: comparison between meningitis and non-meningitis IPD. Vaccine. 2016;34:4543–4550.
12. Ben-Shimol S, Greenberg D, Hazan G, et alIsraeli Bacteremia and Meningitis Active Surveillance Group. Differential impact of pneumococcal conjugate vaccines on bacteremic pneumonia versus other invasive pneumococcal disease. Pediatr Infect Dis J. 2015;34:409–416.
13. Bressan S, Berlese P, Mion T, et alBacteremia in feverish children
presenting to the emergency department: a retrospective study and literature review. Acta Paediatr. 2012;101:271–277.
14. Sáez-Llorens X, McCracken GH Jr.Bacterial meningitis in children
. Lancet. 2003;361:2139–2148.
15. Stoll ML, Rubin LGIncidence of occult bacteremia
among highly febrile young children
in the era of the pneumococcal conjugate vaccine: a study from a Children
’s Hospital Emergency Department and Urgent Care Center. Arch Pediatr Adolesc Med. 2004;158:671–675.
16. Wilkinson M, Bulloch B, Smith MPrevalence of occult bacteremia
aged 3 to 36 months presenting to the emergency department with fever in the postpneumococcal conjugate vaccine era. Acad Emerg Med. 2009;16:220–225.
17. Ribitzky-Eisner H, Minuhin Y, Greenberg D, et alEpidemiologic and microbiologic characteristics of occult bacteremia
among febrile children
in Southern Israel, before and after initiation of the routine antipneumococcal immunization (2005-2012). Pediatr Neonatol. 2016;57:378–384.
18. Central Bureau of Statistics. Statistical Abstracts of Israel, 2014. No. 65. 2014.Jerusalem: State of Israel,
19. Hernandez-Bou S, Trenchs V, Batlle A, et alOccult bacteraemia is uncommon in febrile infants who appear well, and close clinical follow-up is more appropriate than blood tests. Acta Paediatr. 2015;104:e76–e81.
20. Joffe MD, Alpern EROccult pneumococcal bacteremia: a review. Pediatr Emerg Care. 2010;26:448–454; quiz 455.
21. Yildirim I, Little BA, Finkelstein J, et alT; he Massachusetts Dept. of Public Health. Surveillance of pneumococcal colonization and invasive pneumococcal disease reveals shift in prevalent carriage serotypes in Massachusetts’ children
to relatively low invasiveness. Vaccine. 2017;35:4002–4009.
22. Greenberg D, Givon-Lavi N, Faingelernt Y, et alNasopharyngeal pneumococcal carriage during childhood community-acquired alveolar pneumonia: relationship between specific serotypes and coinfecting viruses. J Infect Dis. 2017;215:1111–1116.
23. van Hoek AJ, Sheppard CL, Andrews NJ, et alPneumococcal carriage in children
and adults two years after introduction of the thirteen valent pneumococcal conjugate vaccine in England. Vaccine. 2014;32:4349–4355.
24. Rokney A, Ben-Shimol S, Korenman Z, et alEmergence of Streptococcus pneumoniae serotype 12F after sequential introduction of 7- and 13-valent vaccines, Israel. Emerg Infect Dis. 2018;24:453–461.
25. Ladhani SN, Collins S, Djennad A, et alRapid increase in non-vaccine serotypes causing invasive pneumococcal disease in England and Wales, 2000-17: a prospective national observational cohort study. Lancet Infect Dis. 2018;18:441–451.
26. Camilli R, D’Ambrosio F, Del Grosso M, et alPneumococcal Surveillance Group. Impact of pneumococcal conjugate vaccine (PCV7 and PCV13) on pneumococcal invasive diseases in Italian children
and insight into evolution of pneumococcal population structure. Vaccine. 2017;35:4587–4593.
27. Baraff LJManagement of fever without source in infants and children
. Ann Emerg Med. 2000;36:602–614.
The Israeli pediatric bacteremia and meningitis group
Ron Dagan, principal investigator, Beer Sheva; Jacob Amir, Petah-Tikva; Galia Barkai, Tel Hashomer; Diana Averbuch, Jerusalem; Shraga Aviner, Ashkelon; Ahuva Bachinski, Hadera; Maskit Bar-Meir, Jerusalem; Avihu Bar-Yochai, Zerifin; Ilana Benedikt, Jerusalem; Rita Bernstein, Rehovot; Tal Brosh-Nissimov, Ashdod; Nael Elias, Nazareth; Dan Engelhard, Jerusalem; Moshe Ephros, Haifa; Daniel Glikman, Nahariya; Giora Gottesman, Kfar-Saba; Galia Grisaru-Soen, Tel-Aviv; Alex Guri, Rehovot; Imad Kassis, Haifa; Nathan Keller, Tel Hashomer; Zina Korenman, Jerusalem; Hannah Leskes, Ashdod; Anthony Luder, Safed; Orli Megged, Jerusalem; Dan Miron, Afula; Meirav Mor, Petah-Tikva; Avi Peretz, Tveria; Uri Rubinstein, Netania; Yechiel Schlesinger, Jerusalem; David Schwartz, Tel-Aviv; Itamar Shalit, Petach Tikva; Eli Somekh, Holon; Isaac Srugo, Haifa; Michal Stein, Hadera; Yonatan Yeshayahu, Ashdod; Alvira Zbriger, Hadera; Miriam Zucker, Zfat.