Streptococcus pneumoniae is a leading causative pathogen of common pediatric head and neck infections (HNIs), such as acute otitis media (AOM), meningitis, acute mastoiditis (AM (and acute bacterial sinusitis (ABS). The latter could be accompanied by related complications, for example, orbital cellulitis and subperiosteal abscess.1–4 Before the introduction of pneumococcal conjugate vaccines (PCVs), S. pneumoniae was the cause of ~11% of all deaths in children 1–59 months of age.5 In the last 2 decades, PCVs were integrated worldwide into National Immunization Programs (NIPs) and were found to be effective, safe and immunogenic in infants.6,7
Numerous papers reported on the changes in the epidemiology, microbiology and clinical presentation of pediatric HNIs upon vaccination with PCV7.4,8–11 Of them, AOM and AM were the 2 most commonly investigated infections, because of their relative high prevalence among hospitalized children. Only 1 US study pooled various pediatric HNIs among PCV immunized children aged 1–4 years from the Kids National Inpatient Database (1997–2009); however, it focused mainly on the economic aspects of these infections (hospital stay and healthcare costs).1 To our knowledge, there are limited data on the epidemiology of pediatric HNIs in the post-PCV13 era.
Most pediatric HNIs are managed in an outpatient setting, and since obtaining cultures is routinely nonfeasible, empiric antibiotics treatment is administered. Only few HNIs require hospitalization, when there are suspected complications or if oral antibiotic treatment has failed.12–15 The causative agents of HNIs in nonhospitalized and hospitalized children differ. A recent study demonstrated that children hospitalized with AOM have bacteriologic cultures, which are similar to children with AM, rather than to nonhospitalized children with AOM.16 In this study, we analyzed the incidence and epidemiology of pneumococcal HNIs (pHNIs) in children hospitalized with HNIs in the years prior, during and after the introduction of PCVs in the Israeli NIP.
PATIENTS AND METHODS
Study Design and Population
The study was approved by the local Institutional Review Board. We retrospectively identified children 0–16 years of age, who were hospitalized between January 1, 2007, and December 31, 2014, in the pediatrics department in a secondary hospital (which serves some 150,000 children). There were no appreciable changes in the population size during the study years. Children presenting to the pediatric emergency room with HNIs, who were not hospitalized but eventually had positive pneumococcal cultures as reported by the laboratory a few days after their visit, were excluded, since an analysis on all outcomes could not be performed. Children with congenital malformations of the ear, nasopharynx and palate were excluded, as were children with immunological disorders. For each calendar year, we retrieved the number of admitted patients to the pediatrics department. Rates were calculated per 1000 hospitalized children.
Definition of HNIs
AOM was diagnosed by otoscopy when children presented an acute onset of symptoms (<7 days), consistent with AOM. Diagnosis of AOM was based on the following 2 criteria: (1) middle ear fluid detected by means of an otoscopic examination that was characterized by at least 2 of the following tympanic membrane findings: bulging, decreased or absent mobility (if pneumatic otoscopy was performed), abnormal color or opacity not due to scarring, or air-fluid interfaces and (2) the child presented typical symptoms of fever, otalgia, ear tugging, irritability or loss of appetite.
AM was diagnosed in children presenting typical clinical findings (AOM, postauricular tenderness, erythema or swelling, protruding auricle and palpable/fluctuating mass), associated with systemic symptoms (fever, lethargy, irritability, poor feeding and diarrhea).
ABS diagnosis was in compliance with the diagnostic criteria published by the American Academy of Pediatrics, if (1) upper respiratory tract infection (URI) symptoms had been severe or abrupt, and included concurrent temperature ≥102°F (39°C) and purulent rhinorrhea in an ill-looking child within 3 days of URI onset; (2) when URI symptoms had lasted for >10 days without clinical improvement or (3) when URI symptoms had worsened after an initial improvement.
Meningitis was diagnosed in children presenting an acute onset of headaches, nuchal rigidity, high temperature [≥100.7°F (38.2°C)] and altered mental status that underwent lumbar puncture for cerebrospinal fluid testing.
Definitions of pHNIs
We comprised a patient log of children with pHNIs by retrieving records of children who had (1) relevant International Classification of Diseases-9 codes in their discharge letter (Table, Supplemental Digital Content 1, http://links.lww.com/INF/C603) and (2) pneumococcal culture from the middle ear, blood, abscess, cerebrospinal fluid or had a positive urine pneumococcal antigen. Serotyping for S. pneumoniae was performed only in children presenting suspected meningitis, since it is considered as an “invasive pneumococcal disease” (IPD), which mandates further characterization. No serotyping was available for the other pHNIs, as they are currently not considered as IPDs, and thus do not mandate it. Nevertheless, because of the small number of pneumococcal meningitis cases, serotype data were not included, since any conclusions concerning serotype changes secondary to PCV implementation could not be drawn. We cross-matched and verified that only children who met the eligibility criteria for HNIs were included. For each verified HNI episode, we retrieved the patient’s age, sex and duration of hospitalization. For each verified pHNI episode, we retrieved the patient’s age and sex; history of current disease; culture type and results; duration of hospitalization, the need for imaging studies and surgery and complications. When children were coded for both AOM and AM, they were counted as 1 episode of AM, which is the more severe form of disease.
In Israel, childhood vaccinations against S. pneumoniae are given to all infants at Mother and Child Health Clinics at 2, 4 and 12 months of age. Following its introduction in the Israeli NIP (2009), PCV7 was administered to infants who were born from January 2008. In November 2010, PCV13 replaced PCV7 in the routine NIP. Therefore, 2007–2008 were considered as the “pre-PCV years,” 2009–2011 were considered as the “transition years,” in which both PCV7 and PCV13 were implemented in the NIP (2011 was considered as the “catch-up” year for PCV13), and 2012–2014 were considered as the “post-PCV years.” Because of the short interval between PCV7 and PCV13 introduction, analyzing the changes occurring in the post-PCV7 years was not available; therefore, we focused only on the post-PCV13 years.
Because of the retrospective nature of this study, we could not access immunization data from the patient’s immunization card or other external electronic databases. Yet, our population is quite similar to the one in Southern Israel (situated 60 miles from us), where Ben-Shimol et al17 recently reported a high vaccination rate for PCVs. By June 2011 and December 2012, ~80% and ~90% of 7–11-month-old Israeli children living in Southern Israel received ≥2 PCV7 and/or PCV13 doses, respectively. By June 2014, ~95% of them received ≥2 PCV13 doses. Furthermore, vaccination data were routinely collected by physicians from caregivers upon hospitalization at the pediatrics department in each case. Caregivers were asked if their child’s vaccination status was up-to-date. According to these data, children >4 months of age who received ≥2 doses of PCV7/PCV13 were considered as immunized. All other children were considered as unimmunized.
Data were recorded using Microsoft Excel office software (Microsoft, Redmond, WA). Subjects included were those who met eligibility criteria, who complied with the study design protocol and for whom all data were collected. The statistical analysis was performed per HNI episode, and not per patient. Sequential episodes were considered to be separate events if they were >30 days apart, in order to avoid the bias of inadvertently including partially treated episodes or selecting a special population, for example with an unknown or undiagnosed immunodeficiency. In recurrent cases (>1 admission), episodes were counted twice in the incidence calculation, both in the numerator and the denominator. Statistical analysis was performed using SPSS 17.0 software (IBM Corp., Armonk, NY). Contingency table analysis for comparing rates between unmatched samples was performed using the χ2 test or Fisher exact test, as appropriate. The Student’s independent samples t test was used to compare continuous variables. All tests were considered significant if P values were ≤0.05.
HNIs accounted for 2.5%–4.7% of admissions to the pediatric department in the study years, and there was a downward trend in the incidence in the post-PCV years compared with previous years (Table 1). Boys were more likely to be hospitalized with these infections than girls, 58% (477/820) versus 42% (343/820), P = 0.02. There were no appreciable changes in the average age of hospitalized children and in the average hospitalization duration across the study years.
Figure, Supplemental Digital Content 2, http://links.lww.com/INF/C604, shows a CONSORT diagram on how children were initially identified and eventually included. As shown in Table 2, 87 children were identified with pHNIs: 83 had positive culture and 4 had positive pneumococcal urine antigen (all meningitis patients). AOM (42) followed by AM (28) and meningitis (17) were the most common pHNIs. There were no cases of pneumococcal ABS or pneumococcal head and neck abscesses (ie, Bezold abscess). There were 4 children who had >1 admission: 2 with AOM, 1 with AM and 1 with meningitis. Boys were more likely to have pHNIs than girls, but the difference was statistically insignificant. In the pre-PCV years, there were ~20% positive pneumococcal culture rates, which sharply declined to <10% in the transition years and the post-PCV years. This downward trend was more overt than the observed trend in the overall number of all-cause HNIs and could not be explained by the insignificant changes in the culture acquisition rates during the study period. The average age of children with pHNIs gradually increased in the transition and the post-PCV years, when compared with the pre-PCV years: 1.1 (2007) versus 2.8 years (2014). This was in line with the increasing proportion of unimmunized children presenting with pHNIs during the same periods (data not shown).
The incidence of HNIs and pHNIs per 1000 hospitalized children during the study years, according to the type of infection, is shown in Figure 1A and B, respectively. While there was no substantial change in pHNIs incidence in the pre-PCV years (2007–2008), there was already a decrease in the incidence of HNIs during these years. In 2009 and 2010, there was a sharp decrease (~70%) in the all-cause HNI incidence, mainly due to the sharp decrease of AOM episodes, which contributed ~55% of pHNIs hospitalizations in the pre-PCV years. There were no appreciable changes in the incidence of all-cause AM, ABS and meningitis. Following a temporary increase in 2011, because of a slight surge in pneumococcal AM and ABS incidence, there was an additional decrease in the post-PCV years in the incidence of all pHNIs.
Effect of PCV on Clinical Manifestation
Throughout the study years, children with pHNIs had comparable white blood counts and C-reactive levels upon their admission; the use of imaging studies was overall modest and remained stable; the surgery rate (mostly myringotomies for suppurative AOM and AM, but also few subperiosteal abscess drainages and mastoidectomies) remained almost unchanged (data not shown). Table 3 shows specific comparison between unimmunized and PCV immunized children with pHNIs as previously defined. Unimmunized children were more likely to suffer from pneumococcal infections than immunized children (P = 0.001). Of note, there were no significant differences in the clinical presentation of these 2 groups, except for the need of surgery, which was significantly higher in immunized children (P = 0.042).
The findings of the current study provide further evidence of PCV effectiveness against HNIs in PCV immunized children, and particularly against pHNI which required hospitalization. According to previous studies, the overall incidence of S. pneumoniae infections in children has dramatically declined since PCVs were in common use.2,18 Nevertheless, to the best of our knowledge, no publication documented its benefit in pediatric patients who suffered from HNIs in relation to their immunization status. A substantial reduction in the all-cause HNIs incidence, and more specifically in pHNIs, in PCV immunized children who required hospitalization, was evident in the present study. This was also followed by a significant reduction in AOM rates and to a much lesser extent, in AM and meningitis rates.
The introduction of the Haemophilus influenzae type b vaccine in the 1980s significantly altered the epidemiology and bacteriology of many pediatric infectious diseases, especially HNI.19 The advent of PCVs caused other changes in these common infectious diseases, as previously described.
The all-cause HNIs hospitalization rates described in this study were much higher when compared with data from other countries. For example, all-cause AOM hospitalization rates in Italian children <5 years in 2001–2005 (pre-PCV years) and 2006–2011 (post-PCV7 years) were 20.01 and 10.22/100,000 children/year, respectively.20 When adjusted to ~150,000 children population in our area and stratified to age <5 years, our all-cause AOM hospitalization rates for children <5 years in the pre-PCV years (2007–2008) and post-PCV years (2012–2014) were 58.9 and 26.3/100,000 children/year, respectively. Additionally, the pneumococcal AOM hospitalization rates in these periods were 8.4 and 2.2/100,000 children/year, respectively. These differences could be related to the lax hospitalization policy customary in Israel, especially in infants with URIs and the need to survey them in-house. Another explanation could be the fact that Israeli infants attend daycare centers from a very young age, making viral infections more common, thus enabling HNIs to be more frequent in this population.
We observed a reduction in the HNIs incidence, but not in pHNIs incidence, in the pre-PCV years, due to decrease in AOM hospitalization rates. Because our study population comprised of hospitalized children (and not in the outpatient setting), and data from earlier years (before 2007) were not collected, we cannot draw concrete conclusions. We may attribute this to the publication and implementation of local AOM diagnosis and treatment guidelines published in 2006, which resulted in a decrease in AOM over-diagnosis. A significant reduction in AOM rates, and to a lesser extent, in AM and meningitis rates was observed in the current study. Moreover, a ~50% and ~75% reduction in all-cause AOM and pneumococcal AOM hospitalization rates, respectively, was found. A similar reduction in incidence of meningitis in the post-PCV years was demonstrated. This was found to be in concordance with other reports that showed reduction in all-cause meningitis21 and pneumococcal meningitis22 in the post-PCV period. In a different work from the US, it has been shown that the proportion of pneumococcal isolates from the middle ear and mastoid cultures dramatically decreased in PCV13 immunized children in the first 3 postmarket years of PCV13 (2011–2013), with a sharp decrease in the serotypes contained in the vaccine, and especially serotype 19A.23
PCV immunized children presenting with pHNIs were younger than unimmunized children (but not statistically significant). This observation may reflect that nonvaccine serotypes were more likely to occur in the younger age group, and similarly, the increased likelihood for surgery among PCV immunized children could indicate more severe serotypes affecting those who were immunized. Support for these theories can be found in other papers.8,24,25
Our main strengths are the meticulous collection of data regarding enrolled children, matching the child’s immunization status with the current disease. We demonstrated the effect of both vaccines (PCV7/PCV13) on HNIs, and more specifically on pHNIs epidemiology. Yet, because of the short interval between PCV7 introduction and its replacement by PCV13, within 16 months, we could not show the effect of each vaccine separately. In our view, our results show the bigger net effect of PCV13.
As with any database study, there are limitations of the data itself. Foremost, the use of International Classification of Diseases, 9th revision, Clinical Modification coding is only as accurate as the data coded. Some patients with inaccurate diagnoses may be erroneously included or excluded from this dataset. This limitation was minimized by carefully reviewing retrieved records and assuring that they had HNIs as defined. Furthermore, serotyping data were not available, thus the elimination of vaccine-type serotypes could not be demonstrated, and the presented findings support only the indirect effect of PCVs. However, serotyping is routinely conducted only in IPD” episodes, which were not the majority of cases in this study.
The authors thank Mrs. Nitzhit Zlikovsky, our archivist, for her help in gathering the medical files needed for this study.
1. Walls A, Pierce M, Krishnan N, et alPediatric head and neck complications of Streptococcus pneumoniae
before and after PCV7
vaccination.Otolaryngol Head Neck Surg2015152336–341
2. Esposito S, Principi NImpacts of the 13-valent pneumococcal conjugate vaccine
in children.J Immunol Res20152015591580
3. Pittet LF, Posfay-Barbe KMPneumococcal vaccines for children: a global public health priority.Clin Microbiol Infect201218suppl 525–36
4. Marom T, Tan A, Wilkinson GS, et alTrends in otitis media-related health care use in the United States, 2001-2011.JAMA Pediatr201416868–75
5. O’Brien KL, Wolfson LJ, Watt JP, et alHib and Pneumococcal Global Burden of Disease Study TeamBurden of disease caused by Streptococcus pneumoniae
in children younger than 5 years: global estimates.Lancet2009374893–902
6. Zangwill KM, Greenberg DP, Chiu CY, et alSafety and immunogenicity of a heptavalent pneumococcal conjugate vaccine
7. Vanderkooi OG, Scheifele DW, Girgenti D, et alCanadian PCV13
Study GroupSafety and immunogenicity of a 13-valent pneumococcal conjugate vaccine
in healthy infants and toddlers given with routine pediatric vaccinations in Canada.Pediatr Infect Dis J20123172–77
8. Koutouzis EI, Michos A, Koutouzi FI, et alPneumococcal mastoiditis in children before and after the introduction of conjugate pneumococcal vaccines.Pediatr Infect Dis J201635292–296
9. Kempf M, Varon E, Lepoutre A, et alDecline in antibiotic resistance and changes in the serotype distribution of Streptococcus pneumoniae
isolates from children with acute otitis media
; a 2001-2011 survey by the French Pneumococcal Network.Clin Microbiol Infect20152135–42
10. Halgrimson WR, Chan KH, Abzug MJ, et alIncidence of acute mastoiditis
in Colorado children in the pneumococcal conjugate vaccine
era.Pediatr Infect Dis J201433453–457
11. van der Linden M, Imöhl M, Busse A, et alBacterial spectrum of spontaneously ruptured otitis media in the era of pneumococcal conjugate vaccination in Germany.Eur J Pediatr2015174355–364
12. Grossman Z, Zehavi Y, Leibovitz E, et alSevere acute mastoiditis
admission is not related to delayed antibiotic treatment for antecedent acute otitis media
.Pediatr Infect Dis J201635162–165
13. Zielnik-Jurkiewicz B, Bielicka AAntibiotic resistance of Streptococcus pneumoniae
in children with acute otitis media
treatment failure.Int J Pediatr Otorhinolaryngol2015792129–2133
14. Rosa-Olivares J, Porro A, Rodriguez-Varela M, et alOtitis media: to treat, to refer, to do nothing: a review for the practitioner.Pediatr Rev201536480–486; quiz 487–488
15. Kvaerner KJ, Austeng ME, Abdelnoor MHospitalization for acute otitis media
as a useful marker for disease severity.Pediatr Infect Dis J201332946–949
16. Laulajainen-Hongisto A, Saat R, Lempinen L, et alChildren hospitalized due to acute otitis media
: how does this condition differ from acute mastoiditis
?Int J Pediatr Otorhinolaryngol2015791429–1435
17. Ben-Shimol S, Givon-Lavi N, Leibovitz E, et alNear-elimination 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 Dis2014591724–1732
18. Weil-Olivier C, Gaillat JCan the success of pneumococcal conjugate vaccines for the prevention of pneumococcal diseases in children be extrapolated to adults?Vaccine2014322022–2026
19. Liptak GS, McConnochie KM, Roghmann KJ, et alDecline of pediatric admissions with Haemophilus influenzae
type b in New York state, 1982 through 1993: relation to immunizations.J Pediatr1997130923–930
20. Fortunato F, Martinelli D, Cappelli MG, et alImpact of pneumococcal conjugate universal routine vaccination on pneumococcal disease in Italian children.J Immunol Res20152015206757
21. Tin Tin Htar M, Madhava H, Balmer P, et alA review of the impact of pneumococcal polysaccharide conjugate vaccine (7-valent) on pneumococcal meningitis.Adv Ther201330748–762
22. Regev-Yochay G, Paran Y, Bishara J, et alIAIPD groupEarly impact of PCV7
sequential introduction to the national pediatric immunization plan, on adult invasive pneumococcal disease: a nationwide surveillance study.Vaccine2015331135–1142
23. Kaplan SL, Center KJ, Barson WJ, et alMulticenter surveillance of Streptococcus pneumoniae
isolates from middle ear and mastoid cultures in the 13-valent pneumococcal conjugate vaccine
era.Clin Infect Dis2015601339–1345
24. Ben-Shimol S, Greenberg D, Givon-Lavi N, et alIsrael Bacteremia and Meningitis Active Surveillance Group.Vaccine2016344543–4550
25. Cohen R, Varon E, Doit C, et alA 13-year survey of pneumococcal nasopharyngeal carriage in children with acute otitis media