2 figures, 1 table omitted
Streptococcus pneumoniae (pneumococcus) is
a leading cause of pneumonia and meningitis in the United States and disproportionately
affects young children and the elderly. In 2000, a 7-valent pneumococcal conjugate
vaccine (PCV7) was licensed in the United States for routine use in children
aged <5 years.1 Surveillance data from 2001 and 2002 indicated
substantial declines in invasive pneumococcal disease (IPD) in children and
adults compared with prevaccine years.2,3 This report updates assessment
of the impact of PCV7 on IPD through 2003 by using population-based data from
the Active Bacterial Core surveillance (ABCs) of the Emerging Infections Program
Network, a cooperative surveillance program conducted by several state health
departments and CDC.* The results of this analysis indicated that (1) routine
vaccination of young children with PCV7 continued to result in statistically
significant declines in incidence of IPD through 2003 in the age group targeted
for vaccination and among older children and adults, (2) the vaccine prevented
more than twice as many IPD cases in 2003 through indirect effects on pneumococcal
transmission (i.e., herd immunity) than through its direct effect of protecting
vaccinated children, and (3) increases in disease caused by pneumococcal serotypes
not included in the vaccine (i.e., replacement disease) occurred in certain
populations but were small compared with overall declines in vaccine-serotype
disease. Ongoing surveillance is needed to assess whether reductions in vaccine-serotype
IPD are sustained and whether replacement disease will erode the substantial
benefits of routine vaccination.
ABCs conducted active surveillance for IPD cases through regular contact
with all clinical microbiology laboratories in defined surveillance areas;
periodic audits of laboratory records ensured complete case finding. Pneumococcal
isolates were sent to reference laboratories for serotyping by the quellung
reaction and were categorized as vaccine-type (VT) (serotypes included in
PCV7) or nonvaccine-type (NT) (all other serotypes). A case of IPD was defined
as isolation of pneumococcus from a normally sterile body site (e.g., blood
or cerebrospinal fluid) in an ABCs area resident. Participating areas during
1998-2003 included in this analysis were the state of Connecticut and selected
counties in California, Georgia, Maryland, Minnesota, New York, and Oregon,
representing a total surveillance population of approximately 16 million persons
in 2000. Annual incidence rates were calculated for 1998-1999 by using U.S.
Census Bureau population estimates for those years; incidence rates for 2001-2003
were based on National Center for Health Statistics (NCHS) bridged-race postcensal
population estimates for those years.4 For national projections
of annual numbers of IPD cases, age- and race-specific rates of disease were
applied from the aggregate ABCs surveillance area to the age and racial distribution
of the U.S. population.
The impact of PCV7 introduction on IPD was assessed in three ways. First,
to assess the change in incidence of IPD after PCV7 introduction, IPD rates
for 2001-2003 were compared with the average rate for 1998-1999 (baseline).
Second, the projected number of VT IPD cases directly prevented by PCV7 in
2003 was calculated as the product of (1) the nationally projected number
of VT IPD cases at baseline among children aged <5 years, (2) the 3-dose
coverage of PCV7 in 2003 among all U.S. children aged 19-35 months identified
from National Immunization Survey (NIS) data (68.1%),5 and (3)
vaccine efficacy against VT IPD from a large clinical trial (93.9%).6 Third, the projected number of VT IPD cases indirectly prevented by
PCV7 in 2003 was estimated across all ages aggregately by calculating the
difference between the average annual projected number of VT cases in 1998-1999
and the projected number of VT cases in 2003, and then subtracting the number
of VT cases directly prevented by the vaccine.
From 1998-1999 to 2003, the incidence of VT IPD among children aged
<5 years decreased from 80.0 cases per 100,000 population to 4.6, a decline
of 94% (95% confidence interval [CI] = 92%-96%). The total incidence of IPD
(VT and NT) in this age group declined 75% (CI = 72%-78%), from 96.7 at baseline
to 23.9 in 2003. Incidence rates of VT IPD also declined substantially among
persons outside of the PCV7 target population. For persons aged ≥5 years,
VT disease decreased 62% (CI = 59%-66%) from 1998-1999 to 2003, with the largest
absolute rate reduction occurring among those aged ≥65 years (rate difference:
21.7 cases per 100,000 [rate 33.6 during 1998-1999 and 11.9 during 2003]).
Total IPD incidence declined 29% (CI = 25%-33%), again with the majority of
the absolute rate reduction occurring among those aged ≥65 years (rate
difference: 18.4 cases per 100,000 [rate 60.1 during 1998-1999 and 41.7 during
2003]). The incidence of IPD caused by the 16 serotypes included in the 23-valent
polysaccharide pneumococcal vaccine (PPV23) and not in PCV7 among persons
aged ≥5 years increased 11% (CI = 3%-21%) from 1998-1999 to 2003.
Analysis of the projected 29,599 VT IPD cases prevented nationally by
PCV7 in 2003 compared with 1998-1999 revealed that the majority (69%) of cases
were prevented through indirect effects of the vaccine. An estimated 9,140
cases of VT IPD were directly prevented by vaccinating children aged <5
years with PCV7; an additional 20,459 cases of VT IPD were prevented through
indirect effects of the vaccine across all ages. Incidence of IPD caused by
pneumococcal serotypes not included in PCV7 increased among children aged <5
years and adults aged ≥40 years, with a total of 4,721 projected additional
cases of NT IPD in 2003 compared with the 1998-1999 baseline. After accounting
for this increase, 24,878 net cases of IPD were prevented in 2003; net prevented
cases were evenly distributed between the age group targeted for vaccination
with PCV7 (12,786 prevented cases [51%]) and older children and adults outside
the target population (12,092 prevented cases [49%]).
A Reingold, MD, California Emerging Infections Program, Oakland, California.
J Hadler, MD, Emerging Infections Program, Connecticut Dept of Public Health.
MM Farley, MD, Georgia Emerging Infections Program, Veterans Affairs Medical
Center and Emory Univ School of Medicine, Atlanta, Georgia. L Harrison, MD,
Maryland Emerging Infections Program, Johns Hopkins Bloomberg School of Public
Health, Baltimore, Maryland. R Lynfield, MD, J Besser, MS, Minnesota Dept
of Health. N Bennett, MD, Monroe County Dept of Public Health, Rochester,
New York. A Thomas, MD, Oregon Dept of Human Svcs. W Schaffner, MD, Tennessee
Emerging Infections Program, Vanderbilt Univ Medical Center, Nashville, Tennessee.
B Beall, PhD, Streptococcus Laboratory; T Pilishvili, MPH, Office of Surveillance,
Active Bacterial Core surveillance/Emerging Infections Program Network; CG
Whitney, MD, M Moore, MD, Div of Bacterial and Mycotic Diseases, National
Center for Infectious Diseases; DC Burton, MD, EIS Officer, CDC.
Routine use of PCV7 in young children has reduced the incidence of VT
and overall IPD in children and adults, and these reductions have increased
since 2001.2 The most substantial decline in the rate of VT disease
has been in the target population of children aged <5 years. Data from
2003 also demonstrate statistically significant reductions in the rates of
both VT IPD and total IPD for children aged 5-17 years, whereas no statistically
significant change in disease rate was observed among persons aged 5-19 years
in 2001.2 As of 2003, the total incidence of IPD in persons aged
≥65 years declined to 41.7 cases per 100,000 population in ABCs surveillance
areas, meeting the Healthy People 2010 objective
of no more than 42 cases per 100,000 for this age group.7
Indirect benefits of PCV7 (i.e., cases prevented in unvaccinated persons)
exceeded direct protective benefits among immunized children, with more than
twice as many cases of VT IPD prevented indirectly as directly in 2003. The
indirect effects of PCV7 are believed to be caused by decreased nasopharyngeal
carriage of VT strains among immunized children, which results in decreased
transmission to nonimmunized children and adults (i.e., herd immunity).2,8 On the basis of this mechanism, indirect benefits from PCV7 might
be expected to increase as its vaccination coverage increases. In certain
populations (e.g., children aged <5 years and adults aged ≥40 years),
the reduction in VT IPD attributable to PCV7 was partially offset by an increase
in disease caused by non-VT strains. However, during 2003, the overall magnitude
of this replacement disease was small compared with the reduction in VT disease.
The findings in this report are subject to at least two limitations.
First, secular trends cannot be excluded as a factor in the changing pattern
of IPD in the United States. However, these trends would be expected to affect
disease caused by all serotypes; the reductions in IPD after introduction
of PCV7 have been specific to vaccine serotypes, suggesting a vaccine effect.
The decline in adult IPD likely is not attributable to PPV23, given that no
decline occurred in the incidence of IPD caused by serotypes included in PPV23
but not in PCV7, and given that the slight increase in vaccine coverage of
PPV23 since 19989 would not be expected to cause a measurable change
in IPD rate. Second, the calculations of direct and indirect effects of the
conjugate vaccine were based on data estimates from several sources, each
with an associated margin of error; the calculations in this report provide
only crude estimates of the relative magnitudes of direct and indirect vaccine
effects. In addition, the number of doses of vaccine needed to provide direct
protection is unknown, and partial protection might be provided by fewer than
The robustness of the direct and indirect effects of PCV7 has important
implications for cost-benefit analyses of similar vaccines in the United States
and internationally. Initial estimates of cost-effectiveness for the United
States10 did not account for indirect effects and therefore underestimated
the cost-effectiveness of PCV7. In addition, ongoing surveillance will be
required to monitor the balance of disease reduction versus replacement in
the conjugate vaccine era, particularly in vulnerable populations (e.g., the
elderly and immunocompromised persons), who might be more susceptible to less
virulent non-VT strains of pneumococci. Such information will be critical
for determining whether the composition of conjugate vaccines should be revised
or expanded over time.
This report is based, in part, on contributions by P Daily, MPH, G Rothrock,
MPH, California Emerging Infections Program, Oakland, California. N Barrett,
MPH, Emerging Infections Program, Connecticut Dept of Public Health. W Baughman,
MSPH, J Howgate, MPH, C Payne, MPH, L Rainer, MPH, P Martell-Cleary, MSW,
Georgia Emerging Infections Program, Veterans Affairs Medical Center and Emory
Univ School of Medicine, Atlanta, Georgia. L Sanza, Maryland Emerging Infections
Program, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland.
C Lexau, PhD, R Danila, PhD, B Juni, MS, G Kupferschmidt, Minnesota Dept of
Health. C Long, Univ of Rochester, Rochester, New York; B Anderson, D Hoefer,
New York State Dept of Health. K Stefonek, MPH, Oregon Dept of Human Svcs.
B Barnes, Vanderbilt Univ Medical Center, Nashville; AS Craig, MD, Tennessee
Dept of Health. TH Skoff, MS, ER Zell, MStat, C Wright, Div of Bacterial and
Mycotic Diseases, National Center for Infectious Diseases, CDC.
REFERENCES: 10 available
*Available at http://www.cdc.gov/ncidod/dbmd/abcs.
Direct and Indirect Effects of Routine Vaccination of Children With
7-Valent Pneumococcal Conjugate Vaccine on Incidence of Invasive Pneumococcal
Disease—United States, 1998-2003. JAMA. 2005;294(16):2022–2026. doi:10.1001/jama.294.16.2022