Effectiveness of the US Pneumococcal Conjugate Vaccination Program on Preventing Pneumonia Hospitalizations

On September 28, 2010, Carlos Grijalva made this presentation at the 2010 Annual Conference. Select to access the PowerPoint® presentation (640 KB). Free PowerPoint® Viewer (Plugin Software Help).

Slide 1

Effectiveness of the US pneumococcal conjugate vaccination program on preventing pneumonia hospitalizations

Carlos G. Grijalva, MD MPH
Department of Preventive Medicine
Vanderbilt University School of Medicine
Nashville, Tennessee

Slide 2

Overview

• Pneumococcal diseases:
  • Importance of pneumonia
• PCV7 efficacy against pneumonia
• Study design considerations:
  • PCV7 uptake & secular trends
• Population-based changes after PCV7 introduction

Slide 3

Invasive pneumococcal diseases
Active Bacterial Core Surveillance System, 1999

Image: Bar chart shows the following data:

http://www.cdc.gov/ncidod/dbmd/abcs/survreports/spneu99.pdf

Slide 4

Pneumococcal-related Diseases

Image: A pyramid is shown with the top being the more severe and the bottom more common diseases.

Top of pyramid: Meningitis (More severe), Invasive Pneumococcal Disease (IPD)
Layer 2: Bacteremia (A little less severe than meningitis), also an Invasive Pneumococcal Disease (IPD)
Layer 3: Pneumonia ← Study focus (More common)
Layer 4 (bottom layer): Otitis media/Sinusitis (More common)

Slide 5

Polysaccharide vs. Conjugate vaccines

A table shows the following information:

Adapted from Granoff DM. Vaccines. 2004

Slide 6

Impact of PCV7 on IPD, US

Image: A line graph shows IPD rate / 100,000 by Year. The green line, representing <1, begins just below 150 IPD in 1997 and rises slightly from 1998 to 2000, when it falls sharply to 50 IPD in 2001; this line remains at about 50 IPD until 2009. The red line, representing 1 (PCV7), begins around 175 IPD in 1997 and rises slightly to just above 200 IPD in 1998 and 1999, then falls sharply to ~60 IPD in 2001; this line then remains at about 50 IPD until 2009. The yellow line, representing 2-4, begins at ~40 IPD in 1997 remains at about the same level until 2002, when it falls to ~20 IPD and remains at this level until 2009.

Active Bacterial Core Surveillance (ABCs) Report, Emerging Infections Program Network, Streptococcus pneumoniae, 1997-2009, Centers for Disease Control and Prevention. http://www.cdc.gov/ncidod/dbmd/abcs/survreports.htm. Accessed Sep 16th, 2010.

Slide 7

WHO: Major Causes of Death in Children <5 Years

Image: A pie chart shows the following causes of death:

Pneumonia 19% (This caption is circled in red.)
Injuries 3%
Other 10%
HIV/AIDS 3%
Diarrhea 17%
Measles 4%
Malaria 8%
Neonatal 37%

Image: A bar chart shows the following causes of neonatal deaths:

Other 7%
Tetanus 7%
Diarrhea 3%
Sepsis or pneumonia 26% (This caption is circled in red.)
Asphyxia 23%
Congenital 8%
Preterm 28%

Pneumonia is the leading killer of children.

Bryce J, et al. Lancet 2005;365:1147-1152.

Slide 8

Pneumonia

• Leading infectious cause of death.
• 3% to 18% of all childhood hospitalizations.
• Streptococcus pneumoniae is the leading bacterial cause of pneumonia:
  • 17-44% pneumonia admissions in children.
  • 13-34% pneumonia admissions in adults.

Marston BJ, et al. Arch Intern Med 1997;157:1709-1718.
Farha T, Thomson AH. Paediatr Respir Rev 2005;6:76-82.
Michelow I, et al. Pediatrics 2004;113:701-707.
Drummond P, et al. Arch Dis Child 2000; 83:408-412.
The British Thoracic Society and the Public Health Laboratory Service. Q J Med 1987; 62:195-220.

Slide 9

A table shows the following information: 

Black et al. PIDJ 2002;21:810-15.
Hansen et al. PIDJ 2006;25:779-81.

Slide 10

Objectives

• To estimate the impact of PCV7:
  • Pneumonia hospitalization rates in children aged <2 years (target population).
• To evaluate indirect effects.

Slide 11

Annual No. of PCV7 Doses (millions)
Coverage with 3 or more doses

Image: A bar graph shows the following data:

CDC. Biosurveillance 2000-2005 and National Immunization Survey.
Grijalva CG, et al. Expert Rev Vaccines 2008;7:83-95.

Slide 12

Source of Information: NIS

• HCUP: Nationwide Inpatient Sample
  • Sponsored by AHRQ.
  • Largest inpatient database publicly available.
  • ~20% of US hospital discharges.
  • Discharge level information.
    • De-identified data, diagnoses, procedures, no lab, no chest x-rays.
  • Complex sampling design.

Slide 13

Methods

• Monthly hospitalization rates (annualized):
  • All-cause pneumonia.
  • Pneumococcal pneumonia.
  • Dehydration (control condition).
• Segmented regression analysis: ITS.
• Log-transformed rates as outcomes.
• Quantified vaccine effect by end of 2004.

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 14

Interrupted Time-series Analysis

• Quasi-experimental design.

Wagner AK, et al. J Clin Pharm Therap 2002;27:299-309.

Slide 15

Impact of an Intervention: Before and After Comparison

• Evaluate longitudinal effects of time-delimited interventions:
  • Account for seasonal and secular trends.
• With a control group can assess non-specific changes.

Slide 16

Impact of an Intervention: Before and After Comparison

Image: A bar chart shows the rate per 100,000 before and after an intervention:

1998 (before): 215
2001 (after 2000 intervention): 152

Slide 17

Impact of an Intervention: Before and After Comparison

Image: A bar chart shows the rate per 100,000 before and after an intervention:

1994: 300
1995: 280
1996: 260
1997: 240
1998: 220
1999: 200
2000: 180 (after intervention; a red arrow labeled "Overestimation of Effect" points to this bar.)
2001: 160
2002: 140
2003: 120
2004: 100

Slide 18

Impact of an Intervention: Before and After Comparison

Image: A bar chart shows the rate per 100,000 before and after an intervention:

1998 (before): 180
2001 (after 2000 intervention): 180

Slide 19

Impact of an Intervention: Before and After Comparison

1994: 100
1995: 120
1996: 140
1997: 160
1998: 180
1999: 200
2000: 220 ((after intervention; a red arrow labeled "Overestimation of Effect" points to this bar.)
2001: 240
2002: 260
2003: 280
2004: 300

The bars are highlighted in deep blue up to level 180 for 1998 and 2001.

Slide 20

Pneumococcal Pneumonia Rates/100,000 U.S. Children <2 Years

Image: A chart shows Pre-PCV7 vs. Post-PCV7. Hospitalization rates reach as high as 50 per 100,000 until halfway through 1999 (Pre-PCV7), then begin to fall; in mid-2000, the highest rate is ~25, and ~20 from 2001-2004 (Post-PCV7).

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 21

Pneumococcal Pneumonia Rates/100,000 U.S. Children <2 Years

Image: A chart shows Pre-PCV7 vs. Post-PCV7 is shown. Hospitalization rates reach as high as 50 per 100,000 until halfway through 1999 (Pre-PCV7), then begin to fall; in mid-2000, the highest rate is ~25, and ~20 from 2001-2004 (Post-PCV7). A red line just above 25 shows the average rate of hospitalizations for 1997 through mid-1999; another red line at ~10 shows the average rate of hospitalizations for mid-2000 through 2004.

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 22

Pneumococcal Pneumonia Rates/100,000 U.S. Children <2 Years

Image: A chart shows Pre-PCV7 vs. Post-PCV7. The chart is the same as the one shown in Slide 21, but the red line just above 25 now extends across the chart.

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 23

Pneumococcal Pneumonia Rates/100,000 U.S. Children <2 Years

Image: A chart shows Pre-PCV7 vs. Post-PCV7. The chart is the same as the one shown in Slide 22, but there is now a yellow circle at the 2004 end of the chart with the following text: "-65% (-47, -77)," and a Delta symbol below -77.

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 24

All-cause pneumonia
Rates/100,000 U.S. Children <2 Years

Image: A chart shows Pre-PCV7 vs. Post-PCV7. Hospitalization rates reach as high as 2500 per 100,000 until halfway through 1999 (Pre-PCV7), then begin to fall; the highest rate is ~1800, 2000 in mid-2001 (Post-PCV7), ~1200 in mid-2002, ~1500 in mid-2003, and 1000 in mid-2004. A red line just above 1000 and rising slightly extends across the chart from 1997 through mid-2004; another red line is at ~900 and falls slightly from mid-2000 through 2004. There is a yellow circle at the 2004 end of the chart with the following text: "-39% (-22, -52)," and a Delta symbol below -52.

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 25

Dehydration Rates/100,000 U.S. Children <2 Years

Image: A chart shows Pre-PCV7 vs. Post-PCV7. Hospitalization rates are as high as 1000 per 100,000, with one peak at 1500 in mid-1998 (Pre-PCV7); in mid-2000, the highest rate is ~1500, 1000 in mid-2001 (Post-PCV7), ~1600 in mid-2002, ~1200 in mid-2003, and 500 in mid-2004. A red line just above 500 and rising slightly extends across the chart from 1997 through mid-2004; another red line is at ~900 and falls slightly from mid-2000 through 2004. There is a yellow circle at the 2004 end of the chart with a Delta symbol and -0% above it.

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 26

Pneumococcal Pneumonia Hospitalizations

Image: A bar chart shows the percent of change in pneumococcal pneumonia hospitalizations:

• <2: 62%
• 2-4: 70%
• 5-17: 45%
• 18-39: 30%
• 40-64: 8%
• >64: 22%

Grijalva CG, et al. Expert Rev Vaccines 2008 Feb;7(1):83-95.

Slide 27

All-cause Pneumonia Hospitalizations

Image: A chart shows the percent of change in All-cause Pneumonia Hospitalizations:

• <2: 39%
• 2-4: 17%
• 5-17: 18%
• 18-39: 28%
• 40-64: 20%
• >64: 16%

Grijalva CG, et al. Expert Rev Vaccines 2008 Feb;7(1):83-95.

Slide 28

Estimated Absolute Declines in Pneumonia Hospitalizations

Grijalva CG, et al. Lancet 2007;369:1179-1186.

Slide 29

Impact of PCV7 on Healthcare Use for Pneumonia; U.S. Children <2 Years, 1997-1999 vs 2004

Image: A bar chart shows the percent of change in Impact of PCV7 on Healthcare Use for Pneumonia:

All-cause pneumonia:

• Hospitalizations: 52% (11.5 vs. 5.5 per 1000 person-years)
• Ambulatory: 40% (99.3 vs. 58.5 per 1000 person-years)

Pneumococcal pneumonia:

• Hospitalizations: 58% (0.63 vs. 0.27 per 1000 person-years)
• Ambulatory: 45% (1.73 vs. 0.92 per 1000 person-years)

Zhou F, et al. Arch Pediatr Adolesc Med 2007;161:1162-1168.

Slide 30

Changes in Non-pneumonia ARI hospitalizations, U.S. 1997-2006

Image: A line graph shows the following data:

Grijalva et al, 2009 MMWR 58(1):1-4.

Slide 31

All-cause pneumonia Rates/100,000 U.S. Children <2 Years

Image: A chart shows the the following data:

Grijalva et al. Clin Infect Dis. 2010; 50(6):805-13

Slide 32

Conclusions

• Major declines in all-cause and pneumococcal pneumonia after PCV7 program introduction.
• Sustained declines consistently observed in different studies/settings.
• Large national database (HCUP NIS) allowed detection and monitoring of direct and indirect effects.

Slide 33

Acknowledgement

• Marie R. Griffin, MD MPH
  • Professor of Medicine and Preventive Medicine. VUMC
• J. Pekka Nuorti, MD DSc
  • Epidemiologist, CDC

Images: Photographs of the presenters are shown.