Secondary Logo

Journal Logo

Original Studies

Population-Based Surveillance for Childhood Invasive Pneumococcal Disease in the Era of Conjugate Vaccine

Hsu, Katherine MD*; Pelton, Stephen MD*; Karumuri, Sudharani MPH; Heisey-Grove, Dawn MPH; Klein, Jerome MD* Massachusetts Department of Public Health Epidemiologists

Author Information
The Pediatric Infectious Disease Journal: January 2005 - Volume 24 - Issue 1 - p 17-23
doi: 10.1097/01.inf.0000148891.32134.36
  • Free


Prevention of pneumococcal pneumonia by immunization with specific capsular polysaccharides was first demonstrated in U.S. military recruits in the early 1940s.1 Interest in pneumococcal vaccine research diminished with succeeding decades but was revived in large part because of concern for emerging resistance of Streptococcus pneumoniae to antimicrobial regimens. Use of pneumococcal polysaccharide vaccines in pediatrics was limited because polysaccharide antigens were insufficiently immunogenic in the target age group with highest risk for pneumococcal disease (children younger than 2 years of age). However, a rapid decline in invasive disease caused by Haemophilus influenzae type b after the introduction of H. influenzae type b conjugate vaccines in North America and Western Europe provided proof of concept that conjugation of polysaccharide antigen to carrier protein elicits protective antibody in infants and toddlers.

A 7-valent pneumococcal conjugate vaccine containing 2 μg each of saccharides of serotypes 4, 9V, 14, 18C, 19F and 23F and 4 μg of 6B, linked to the protein carrier CRM197, a nontoxic mutant of diphtheria toxin, was licensed in February 2000 in the United States (PCV7, Prevnar; Wyeth). PCV7 was recommended for universal administration to children younger than 2 years and high risk children 2–5 years of age and was also recommended for consideration for moderate risk children 2–5 years of age by the Advisory Committee for Immunization Practices,2 American Academy of Pediatrics3 and American Academy of Family Practitioners.4 Children deemed to be at high risk for invasive pneumococcal infection included children with sickle-cell disease or other hemoglobinopathies, congenital or acquired asplenia or splenic dysfunction, chronic cardiac and pulmonary disease (excluding asthma), diabetes mellitus, cerebrospinal fluid leaks and human immunodeficiency virus (HIV) infection or other immunocompromising conditions such as malignancies, chronic renal failure or nephrotic syndrome, receipt of immunosuppressants including long term corticosteroids or receipt of a solid organ transplant. Children deemed to be at moderate risk for invasive pneumococcal disease included children 24–35 months of age; children 36–59 months of age of Alaskan Native, American Indian or African American descent; and children 36–59 months of age who attended group day-care centers. The standard U.S. vaccination schedule for infants 2–6 months old was a 3-dose primary series followed by a booster at age 12–15 months. Full scale distribution of PCV7 per guidelines began in July 2000 in Massachusetts, where all recommended childhood vaccines are purchased by the state and distributed by the Massachusetts Department of Public Health to both public and private health care providers. From January 2002 through May 2003, a moderate PCV7 shortage existed in Massachusetts, and providers were asked to defer a fourth dose of vaccine in healthy infants as well as the single dose of vaccine considered for healthy toddlers older than 24 months.5

In October 2001, we initiated population-based enhanced passive surveillance for invasive pneumococcal disease (IPD) in children younger than 18 years to evaluate the impact of universal pneumococcal conjugate vaccination on childhood IPD in Massachusetts. This report details IPD incidence rates during the 2-year period October 2001 through September 2003 in children in Massachusetts and examines risk features associated with children who suffered invasive pneumococcal disease.


A case of childhood invasive pneumococcal disease was defined by a positive culture for S. pneumoniae from normally sterile body fluid from a Massachusetts resident younger than 18 years of age. Pneumococcal meningitis was defined as either a clinical diagnosis of meningitis by the hospital record as reported by the primary care provider or a positive culture from cerebrospinal fluid. Clinical microbiology laboratories in Massachusetts were requested to report and send in all invasive S. pneumoniae isolates to the Massachusetts Department of Public Health. Epidemiologists from the health department periodically contacted and visited microbiology laboratories to encourage complete reporting. State health department epidemiologists used a standardized data collection form to interview the case's primary care provider to determine the age, gender, race, vaccination status, clinical presentation and underlying medical conditions of the case. Parent/guardians were interviewed to confirm race and identify day-care attendance and household information such as occupancy number. Chart review was not performed on cases. Serotyping was performed on available isolates at the Maxwell Finland Laboratory for Infectious Diseases at Boston Medical Center. Serotyping via quellung reaction was performed with pneumococcal antisera (Danish Statens Serum Institute, Copenhagen, Denmark).

We report on data from our first 24 months of surveillance through September 2003. Annual incidence rates were derived based on IPD case numbers as numerator values and U.S. Census Bureau estimates from 2000 as denominator values. Annual incidence rates were compared with Massachusetts data collected and previously reported with the use of comparable methods in 1990–1991.6 Specific serogroup information from 1990–1991 was available only for nonmeningitis invasive pneumococcal disease. Statistical analyses were conducted with Microsoft Access 2000 and SAS version 8.0. Ninety-five percent confidence intervals were calculated, and 2-sided P values of <0.05 were considered statistically significant. χ2 tests were used to compare proportions.

We defined cases to be completely vaccinated if they had received (1) 3 PCV7 doses if first dose occurred before age 12 months, (2) 2 PCV7 doses if first dose occurred at age 12–23 months or (3) 1 PCV7 dose if first dose occurred at age 24 months or older; and if last dose of vaccine occurred >2 weeks before disease. Cases were assessed as incompletely vaccinated if at least 1 PCV7 dose had been given >2 weeks before disease but the case did not reach criteria for complete vaccination. Cases were assessed as not vaccinated if either they had no history of PCV7 vaccination or disease occurred within 2 weeks of a single PCV7 dose.

Vaccine failure was defined as vaccine serotype disease in a completely vaccinated case. Vaccine-related failure was defined as vaccine-related serotype disease (eg, 6A, 19A, 23A, etc.) in a completely vaccinated case. Nonvaccine serogroup disease was defined as disease caused by a serotype neither in nor related to vaccine.

Potential vaccine shortage cases were defined as a subset of vaccine serogroup cases occurring after December 28, 2001 when moderate shortage was announced in Massachusetts. The potential vaccine shortage case also had to have had the primary PCV7 immunization series, but no booster dose administered between 12–15 months. Once a potential vaccine shortage case was identified, the patient's primary care provider was also contacted by a state health department epidemiologist to confirm that the booster dose had not been administered because of vaccine shortage.

We had approval from Institutional Review Boards of Boston University Medical Center and Massachusetts Department of Public Health for this study.


Decline in Invasive Pneumococcal Disease.

A total of 191 culture-positive pediatric cases of IPD were identified between October 1, 2001 and September 30, 2003. One hundred thirty-eight cases (72%) occurred in children younger than 5 years of age; 101 of 138 cases (73%) were available for serotyping. The annual incidence rate for invasive disease was 17.4 per 100,000 children younger than 5 years of age in the combined 2001–2003 time period (Table 1), compared with 56.9 per 100,000 per year in the combined 1990–1991 time period.6 Nonmeningitis vaccine serogroup (VSG) disease declined in 2001–2003 compared with 1990–1991; declines were also statistically significant when comparisons were made for each of the 7 serogroups related to the vaccine (Table 1). In children younger than 5 years of age, the reported decline in meningitis was comparable with that observed for nonmeningitis disease.

Changes in Annual Incidence Rates of IPD Among Children Younger Than 5 of Age, According to Year and Serogroup†

Serotype Distribution, 2001–2003.

Figure 1 summarizes specific serotypes causing IPD in children younger than 18 years during the study period. Among 136 isolates (71%) available for typing, 19A and 19F were the most common serotypes identified (each 10%), closely followed by serotype 6A (9%), and nonvaccine serogroups 7 and 15 (each 8%). Among 61 fully vaccinated children younger than 5 years of age with IPD, nonvaccine serogroup (NVSG) disease was more common (37 cases, 61%); but among 38 incompletely vaccinated or unvaccinated children younger than 5 years of age, VSG disease represented 50% of cases (Table 2). In older children, NVSG disease likewise represented a smaller proportion of cases that were incompletely vaccinated or unvaccinated, compared with children who were fully vaccinated. Over all age categories, one-half of the isolates were VSG and one-half were NVSG. In the first 2 years of surveillance, a borderline significant increase in nonmeningitis NVSG disease was seen in children younger than 5 years compared with 1990–1991 (change in annual incidence rate, +99%; P = 0.04; Table 1).

IPD serotypes among children younger than 18 years of age, 2001–2003. Serotype available for 136 of 191 cases in children younger than 18 years of age. Vacc.-rel. indicates vaccine-related.
Serotype Distribution of Invasive Pneumococcal Disease by Age and Vaccine Status, 2001–2003

Risk Features for IPD in Children Younger Than 18 Years of Age During 2001–2003: Age, Race and Comorbid Features.

The majority of IPD cases and peak annual incidence rate of IPD occurred in the 0- to 11-month age group (36.5 cases per 100,000; Fig. 2), followed by the 12–23 months age group (25.3 cases per 100,000). Rates were 8.5 per 100,000 in those 24–59 months of age and 2.4 per 100,000 in those from 60 months up to 18 years of age. Twenty-three cases (12%) of IPD occurred in children younger than 6 months of age; 8 of those cases were VSG disease. VSG disease comprised the largest proportion of IPD in all age groups, except in children younger than 6 months and children 12–23 months of age.

Annual incidence rates of IPD by age and serogroup, 2001–2003. Incidence rates calculated based on U.S. Census 2000 data for Massachusetts.9 Serogroup and age available for 136 of 191 cases.

There was a male predominance among cases. Annual incidence rate was 7.2 per 100,000 male population younger than 18 years, versus 5.5 per 100,000 female population younger than 18 years (relative risk, 1.3; P = 0.08).

The annual incidence rate of IPD in white children younger than 18 years of age was 4.9 per 100,000 (Table 3). Annual incidence rates were significantly higher in black and Hispanic children (respectively, 11.3 and 9.2 per 100,000). Asian/Pacific Islanders likewise experienced a higher rate of disease (5.9 per 100,000) compared with white children, but relative risk of IPD compared with white children was not significantly increased. Race/ethnicity data were missing for 17 cases.

Annual Incidence Rates of IPD by Race Among Children Younger Than 18 Years of Age, 2001–2003

Fifty-nine cases of IPD in children younger than 18 years of age were reported to have comorbid conditions. Table 4 shows the distribution of IPD cases in children with comorbid conditions by age. Traditional comorbid conditions already recognized as risk features for IPD were found in 30 (16%) and were equally distributed above and below 5 years of age. Nontraditional comorbidities such as metabolic disorders, asthma and prematurity were found in 29 (15%) cases and were observed most frequently in children younger than 5 years of age.

Distribution of Comorbid Features by Age at IPD, 2001–2003

Vaccine Serogroup Disease in PCV7-Immunized Children.

Cases representing potential vaccine failures are presented in Table 5. There were 11 cases of vaccine failure and 15 cases of vaccine-related failure. Nine (35%) of 26 cases occurred in children with traditional comorbid risk features for IPD. Among 110 cases of nonvaccine failure IPD (either because the case was caused by a nonvaccine serogroup strain or because the case was not completely immunized), 19 (17%) occurred in children with traditional comorbid conditions.

Vaccine Failures, Vaccine-Related Failures and Potential Vaccine Shortage Cases, 2001–2003†

We also identified 4 IPD cases potentially attributable to vaccine shortage (Table 5). One of these cases was caused by serotype 18C, a vaccine serotype; the other 3 were vaccine-related serotype disease (6A and 2 cases of 19A).


Compared with a decade ago, there has been a 69% reduction in IPD in children younger than 5 years of age in Massachusetts. Similar reductions have been observed in 2 other population-based studies after the licensure of PCV7 for use in the United States, one in the Northern California Kaiser Permanente population12 and the other based on a national surveillance network of selected counties in 8 states.13 The reduction in Massachusetts occurred despite vaccine shortage and with vaccine penetrance in Massachusetts reported as 62 ± 5.5% of children 19–35 months of age having received ≥3 doses of PCV7.14

Comparisons with the 1990–1991 baseline, IPD incidence and typing data were limited for 4 reasons: (1) only serogrouping, not serotyping, was available in 1990–1991; (2) specific serogroup information was only available for nonmeningitis invasive pneumococcal disease; (3) surveillance was performed only on children younger than 5 years of age; and (4) completeness of reporting in either surveillance time period is unknown. For example, it is possible that an increase in completeness of reporting in 2001–2003 could explain the apparent increase in nonmeningitis nonvaccine serogroup disease over the course of the decade. Nevertheless we noted a shift in proportion of disease caused by VSG in children younger than 5 years of age. During this early period of PCV7 usage, about one-half of IPD in children younger than 5 years was caused by VSG disease, versus >80% in the years before PCV7 usage.6,15,16 It is likely that serogroup shifts in disease occurred even before conjugate vaccine usage, as suggested by data from Boston Medical Center where the proportion of bacteremia caused by VSG declined from 91.9% to 80.0% between 1989 and 1998.16 Although almost one-third of isolates were unavailable for serotyping (representing isolates that failed to grow on arrival at the referral laboratory), we have no expectation that serotype distribution of missing isolates was different than available isolates.

Analysis of current (2001–2003) cases of IPD revealed that ∼12% of IPD occurred in children younger than 6 months of age, before completion of the primary series. To have an impact on disease reduction in this subgroup of children, we must rely on herd immunity (decreased colonization with vaccine serotypes resulting in decreased transmission to this population), immunization with a vaccine with broader serotype coverage, strategies involving maternal immunization (with transplacental transfer of antibodies) or earlier age of immunization (a schedule such as that used in a South African vaccine trial of nonavalent pneumococcal conjugate vaccine (6, 10, 14 weeks of age)17).

Black and Hispanic children had significantly higher disease rates than did white children. Higher attack rates were observed in these cohorts for both vaccine and nonvaccine serogroup disease, strongly suggesting increased risk unrelated to access to vaccine. Massachusetts PCV7 vaccine penetrance data also indicated that 62 ± 5.5% of children 19–35 months of age received ≥3 doses of PCV7; although specific data were not available for black and Hispanic children, it is known that 59.0 ± 6.8% of white non-Hispanic children 19–35 months of age received ≥3 doses of PCV7.14 A higher risk for IPD is consistent with previous reports of increased rates of pneumonia and overall IPD in minority populations,13,18 but recent data suggest that the gap in incidence of IPD between blacks and whites may be narrowing.19 We are unable to comment on whether racial disparities observed in the 2001–2003 cohort were reduced compared with 1990–1991, because race data were not reported in the earlier study.

A spectrum of comorbid conditions was identified in our cases from 2001–2003. These included atopy (asthma, eczema), idiopathic thrombocytopenic purpura, gastroesophageal reflux disease, autism, seizure disorders, prematurity, quadriplegia and metabolic disease, in addition to conditions previously recognized as risk features for pneumococcal disease or its complications. Traditional risk features such as sickle-cell disease, HIV, congenital immunodeficiency and malignancy were equally divided among children younger and older than 5 years of age consistent with the lifelong risk for IPD in children with these conditions. This study is unable to identify whether immunization with PCV7 provides protection against IPD in children with such risk features; however, a report from South Africa demonstrated decreased efficacy of PCV in HIV-infected children,17 and certainly those with congenital immunodeficiency may be unlikely to develop protective antibody after immunization. More comorbid conditions such as asthma, metabolic disorders and prematurity occurred in children younger than 5 years of age during the time period of greatest risk for IPD in all children. It is not possible to determine whether any of these conditions increased the risk of IPD because the prevalence of each in the study population was not accurately known. However, prematurity has recently been associated with an increased relative risk for IPD.20

Only a small proportion of observed disease was attributable to vaccine failure or vaccine shortage. Of 191 cases, 11 (6%) were VST failures, 15 (8%) were VRST failures and 4 (2%) were possibly the result of vaccine shortage. Most of these vaccine failure or shortage cases were caused by serotypes 19F or 19A. Efficacy against 19F or 19A has already been reported as lower than other vaccine and vaccine-related serotypes.13,21,22

The interpretation of declining rates of invasive disease may be confounded by whether clinical strategies for assessment of febrile infants have changed with the introduction of PCV7. A survey of Massachusetts pediatric clinicians conducted 1 year after PCV7 licensure found evidence of a 9–31% decrease in the likelihood of obtaining laboratory evaluations (complete blood count, blood culture, urine analysis and culture) for a well-appearing febrile infant with a prior history of 3 doses of PCV7.23 However, one might expect that that provider practices toward diagnosing meningitis would not have been affected and rates of decline for meningitis and nonmeningitis disease are similar, suggesting the observed decline is not simply an artifact of changing practice. A second limitation is the passive nature of reporting of cases of IPD in this study. Although contact with each microbiology laboratory was made multiple times during the 2 years, only isolates from cases of bacterial meningitis were required reporting by Massachusetts statute until February 2003 when all isolates of S. pneumoniae from sterile body fluids became required.

We have identified that infants younger than 1 year of age and minority ethnic groups are at greatest risk for IPD in the era of conjugate vaccine. In addition, almost 40% of cases in children older than 5 years of age were associated with recognized comorbid conditions such as HIV, congenital immunodeficiency, sickle-cell disease and malignancy. Febrile children with these risk features continue to warrant vigilance for pneumococcal disease even if immunized with PCV7. Further reductions in pneumococcal disease in the United States can potentially be achieved by earlier immunization with PCV7, addition of serogroups such as 7 and 15 to current vaccine and improved protection against serotypes 19F and 19A.


We thank the many Massachusetts pediatric health care providers and state health department epidemiologists involved in this study for providing care information. We also thank Drs Dagna Laufer and Frank Malinoski at Wyeth for support and encouragement; and Stephanie Schauer, Kristin Sullivan, Dr Susan Lett and Dr Alfred DeMaria Jr. from the Massachusetts Department of Public Health for their advice and editorial assistance.


1. MacLeod CM, Hodges RG, Heidelberg M, Bernhard WG. Prevention of pneumococcal pneumonia by immunization with specific capsular polysaccharides. J Exp Med. 1945;82:445–465.
2. Centers for Disease Control and Prevention. Preventing pneumococcal disease among infants and children: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR. 2000;49(RR-9):1–35.
3. American Academy of Pediatrics Committee on Infectious Diseases. Policy statement: recommendations for the prevention of pneumococcal infections, including the use of pneumococcal conjugate vaccine (Prevnar), pneumococcal polysaccharide vaccine and antibiotic prophylaxis. Pediatrics. 2000;106(2, pt 1):362–366.
4. American Academy of Family Physicians. Pneumococcal conjugate immunization: April 27, 2000. Available at: Accessed June 4, 2003.
5. Centers for Disease Control and Prevention Advisory Committee on Immunization Practices. Updated recommendations on the use of pneumococcal conjugate vaccine in a setting of vaccine shortage. MMWR. 2001;50:1140–1142.
6. Loughlin AM, Marchant CD, Lett SM. The changing epidemiology of invasive bacterial infections in Massachusetts children, 1984 through 1991. Am J Public Health. 1995;85:392–394.
7. U.S. Census. 1990 data for Massachusetts children younger than 5 years old. Available at: Accessed February 8, 2004.
    8. U.S. Census. 2000 data for Massachusetts children younger than 5 years old. Available at: Accessed February 8, 2004.
      9. U.S. Census. 2000 data for Massachusetts. Available at: Accessed February 8, 2004.
        10. U.S. Census. 2000 data for Massachusetts. Available at: Accessed February 8, 2004.
          11. U.S. Census. 2000 data for Massachusetts. Available at: Accessed February 8, 2004.
            12. Black SB, Shinefeld HR, Hansen J, Elvin L, Laufer D, Malinoski F. Postlicensure evaluation of the effectiveness of seven valent pneumococcal conjugate vaccine. Pediatr Infect Dis J. 2001;20:1105–1107.
            13. Whitney CG, Farley MM, Hadler J, et al. Decline in invasive pneumococcal disease after the introduction of protein-polysaccharide conjugate vaccine. N Engl J Med. 2003;348:1737–1746.
            14. National Immunization Survey. Estimated vaccination coverage with individual vaccines and selected vaccination series among children 19 to 35 months of age by state, 2002. Available at: Accessed January 2, 2004.
            15. Hausdorff WP, Bryant J, Kloek C, Paradiso PR, Siber GR. The contribution of specific pneumococcal serogroups to different disease manifestations: implications for conjugate vaccine formulation and use, part II. Clin Infect Dis. 2000;30:122–140.
            16. Babl FE, Pelton SI, Theodore S, Klein JO. Constancy of distribution of serogroups of invasive pneumococcal isolates among children: experience during 4 decades. Clin Infect Dis. 2001;32:1155–1161.
            17. Klugman KP, Madhi SA, Huebner RE, Kohberger R, Mbelle N, Pierce N. Trial of 9-valent pneumococcal conjugate vaccine in children with and those without HIV infection. N Engl J Med. 2003;349:1341–1348.
            18. Black SB, Shinefeld HR, Ling S, et al. Effectiveness of heptavalent conjugate vaccine in children younger than five years of age for prevention of pneumonia. Pediatr Infect Dis J. 2002;21:810–815.
            19. Flannery B, Schrag S, Bennett NM, et al. Impact of childhood vaccination on racial disparities in invasive Streptococcus pneumoniae infections. JAMA. 2004;291:2197–2203.
            20. Shinefeld H, Black S, Ray P, Fireman B, Schwalbe J, Lewis E. Efficacy, immunogenicity, and safety of heptavalent pneumococcal conjugate vaccine in low birth weight and preterm infants. Pediatr Infect Dis J. 2002;21:182–186.
            21. Park SY, VanBeneden CA, Martin MT, Pilishvili T, Facklam R, Whitney CG. Tracking pneumococcal disease in children receiving the pneumococcal conjugate vaccine. In: 41st Annual Meeting of the Infectious Diseases Society of America, October 9–12, 2003. San Diego, CA. Alexandria, VA: Infectious Diseases Society of America, 2003; Abstract 910.
            22. Kaplan SL, Mason EO, Wald ER, et al. Decrease of invasive pneumococcal infections in children among 8 children's hospitals in the United States following the introduction of the 7-valent pneumococcal conjugate vaccine. Pediatrics. 2004;113:443–449.
            23. Lee KC, Finkelstein JA, Miroshnik Z, et al. Pediatricians’self-reported clinical practices and adherence to national immunization guidelines after the introduction of pneumococcal conjugate vaccine. Arch Pediatr Adolesc Med 2004;158,695–701.
            24. Hsu KK, Pelton SI, Heisey-Grove DM, Hashani S, Klein JD. Sarotype-specific saruoilante for childhood invasive pneumococcal disease. In: 42nd Interscience Conference for Antimicrobial Agents and Chemotherapy, September 27–30, 2002, San Diego, CA. Washington, DC: American Society for Microbiology, 2002. Abstract G-1067.
              25. Hsu KK, Pelton SI, Heisey-Grove DM, et al. Risk factors for childhood invasive pneumococcal disease in PCV7 era. In: Pediatric Academic Societies’ Annual Meeting, May 3–6, 2003, Seattle, WA. Woodlands, TX: Pediatric Academic Societies, 2003. Abstract 1758.

                childhood invasive pneumococcal infections; conjugate vaccines; population surveillance; serotyping

                © 2005 Lippincott Williams & Wilkins, Inc.