ESTIMATING THE BURDEN OF PNEUMOCOCCAL PNEUMONIA AMONG ADULTS Maria A. Said1, Hope L. Johnson2, Bareng A.S. Nonyane2, Maria Deloria-­‐Knoll2, Katherine L. O’Brien2, and AGEDD Adult Pneumococcal Burden Team3 1. Division of InfecPous Diseases, Johns Hopkins School of Medicine, BalPmore, Maryland, USA. 2 IVAC, Department of InternaPonal Health, Johns Hopkins Bloomberg School of Public Health, BalPmore, Maryland, USA. 3. Members and AffiliaPons listed at end of paper INTRODUCTION •  Community-­‐acquired pneumonia (CAP) causes significant morbidity and mortality and a large propor(on is poten(ally aZributable to pneumococcus.1 •  In 2000, in children <5 years old, an es(mated 14 million cases and 750,000 deaths from pneumococcal pneumonia occurred; 2 however, the adult burden is not well characterized. •  Diagnos(c tests for non-­‐bacteremic pneumococcal pneumonia are limited; thus, most e(ology studies measure only the incidence of invasive or bacteremic disease. •  Through a comprehensive review and meta-­‐analysis of studies u(lizing the BinaxNOW® S. pneumoniae urinary an(gen test (UAT), blood cultures, and sputum cultures, we sought to beZer understand the rela(onships between bacteremic pneumococcal pneumonia, non-­‐
bacteremic pneumococcal pneumonia, and all-­‐cause CAP. •  The inten(on is to apply these parameters to a comprehensive review of bacteremic pneumococcal disease among adults in order to project the global disease burden in adults. •  The majority of studies came from resource-­‐rich countries. •  The es(mated ra(o of non-­‐bacteremic to bacteremic pneumococcal pneumonia was 3.75 (95% CI: 2.51-­‐5.59). •  The propor(on of pneumococcal pneumonia that was bacteremic to be 24.7% (95% CI: 21.2%-­‐28.7%) (Figure 3), providing a ra(o of non-­‐bacteremic to bacteremic pneumococcal pneumonia of 3.05 (95% CI: 2.48-­‐3.72). Figure 3. Forest Plot for the Propor(on of Pneumococcal Pneumonia Iden(fied as Bacteremic Study
ID
Author
Proportion (95% CI)
Marekovic et al. (n=80)
1
Krcová et al. (n=84)
2
Hohenthal et al. (n=384)
4
Lasocki et al. (n=108)
5
Endeman et al. (n=201)
6
Snijders et al. (n=213)
7
Van der Eerden et al. (n=262)
8
Hernes et al. (n=20)
9
Beovic et al. (n=109)
11
Andreo et al. (n=107)
12
Falguera et al. (n=177)
14
Fernández-Sabé et al. (n=1474) 15
Gutiérrez et al. (n=493)
16
Marcos et al. (n=398)
17
Ortega et al. (n=128)
18
Perelló et al. (n=64)
19
Sordé et al. (n=474)
20
Johansson et al. (n=184)
22
Strålin et al. (n=235)
23
Kobashi et al. (n=156)
26
Lauderdale et al. (n=168)
27
Charles et al. (n=885)
28
Weatherall et al. (n=59)
29
Murdoch et al. (n=474)
30
Díaz et al. (n=176)
31
Butler et al. (n=149)
33
Nuermberger et al. (n=487)
34
Watt et al. (n=68)
35
Overall (I-squared = 94.1%, p = 0.000)
METHODS Strategy •  We performed a comprehensive literature search to iden(fy papers that compare the diagnos(c yield of blood cultures, sputum cultures, and the Binax UAT. •  We contacted authors of poten(ally relevant studies to obtain informa(on on the numbers of pa(ents undergoing each diagnos(c test, the numbers of posi(ve test results, and the overlaps in posi(ve results among the three methods of detec(on (Figure 1). Figure 1. The rela(onship in diagnos(c test yield of blood culture, sputum culture, and the Binax S. pneumoniae urine an(gen test 0.02
0.17
0.49
0.16
0.29
0.35
0.36
0.30
0.09
0.21
0.31
0.44
0.12
0.32
0.34
0.29
0.31
0.51
0.33
0.07
0.13
0.27
0.05
0.14
0.35
0.17
0.30
0.27
0.25
(0.00, 0.09)
(0.10, 0.27)
(0.44, 0.54)
(0.10, 0.24)
(0.23, 0.36)
(0.29, 0.42)
(0.31, 0.43)
(0.15, 0.59)
(0.05, 0.17)
(0.15, 0.30)
(0.25, 0.39)
(0.42, 0.47)
(0.09, 0.15)
(0.27, 0.37)
(0.26, 0.43)
(0.20, 0.43)
(0.27, 0.35)
(0.44, 0.59)
(0.28, 0.40)
(0.04, 0.13)
(0.09, 0.20)
(0.24, 0.30)
(0.02, 0.15)
(0.11, 0.18)
(0.29, 0.43)
(0.12, 0.24)
(0.27, 0.35)
(0.18, 0.40)
(0.21, 0.29)
NOTE: Weights are from random effects analysis
.1
•  The es(mated propor(on of CAP aZributable to pneumococcus was 27.5% (95% CI: 24.1%-­‐31.3%). The propor(on of CAP iden(fied as pneumococcus varied by diagnos(c test. It was smallest with blood cultures and greatest with the Binax UAT (Figure 4). •  The Binax UAT diagnosed an addi(onal 11.3% (95%CI: 9.5-­‐13.4%) of CAP beyond that iden(fied by blood and sputum culture alone Figure 4. ProporPon of CAP aWributable to Pneumococcus .2
StaPsPcal Analysis 2
.6
•  The total CAP popula(on (N) included only those with radiographic pneumonia. •  Pneumococcal pneumonia: CAP and at least one of the three posi(ve laboratory tests. •  Bacteremic pneumococcal pneumonia: CAP and a blood culture posi(ve for pneumococcus (A/N). •  Non-­‐bacteremic pneumococcal pneumonia: CAP and a sputum culture posi(ve for pneumococcus and/or a posi(ve Binax UAT, without a blood culture posi(ve for pneumococcus: (E+F+I)/N •  The propor(on of CAP aZributable to pneumococcus: (A+E+I+F)/N •  The ra(o of non-­‐bacteremic to bacteremic pneumococcal pneumonia: (E+I+F)/A •  The addi(onal contribu(on of the Binax UAT to blood and sputum culture: (E/N) •  The propor(on of pneumococcal pneumonia that is bacteremic: A/(A+E+I+F) 1
.4
DefiniPons and Outcome Measures .25 .5
0
•  The propor(ons were es(mated using a random effects meta-­‐analysis. •  The ra(o of non-­‐bacteremic to bacteremic pneumococcal pneumonia was es(mated using a nonparametric bootstrap approach. RESULTS 1
2
3
4
5
6
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
Blood/CAP
UAT/CAP
•  A comprehensive literature search and hand searching of cita(on lists from other relevant published studies yielded 469 ar(cles; 35 were ul(mately included in the analysis (Figure 2). Figure 2. Flow Diagram for the SelecPon of Studies 7
Sputum/CAP
Blood+UAT+Sputum/CAP
CONCLUSION •  For every case of bacteremic pneumococcal pneumonia, there are at least 3 addi(onal cases of non-­‐
bacteremic pneumococcal pneumonia. This result was observed using two unique sta(s(cal methods. •  Addi(onal studies from resource-­‐poor countries would be helpful in assuring the validity of this result in these areas. •  This ra(o can be applied to global es(mates of invasive pneumococcal disease or bacteremic pneumococcal pneumonia in adults in order to beZer es(mate the overall burden of disease. •  At least one-­‐quarter of CAP cases are likely aZributable to pneumococcus. •  The Binax UAT may iden(fy an addi(onal 11% of CAP cases with an e(ology of pneumococcal pneumonia. This might be considered in e(ology studies among adults in the resource-­‐limited countries in which only blood cultures and sputum studies are u(lized. REFERENCES •  Fine MJ, Smith MH, Carson CA, et al. Prognosis and outcomes of pa(ents with community-­‐
acquired pneumonia: a meta-­‐analysis. JAMA 1995;274:134-­‐141. •  O'Brien KL, Wolfson LJ, WaZ JP, et al. Burden of disease caused by Streptococcus pneumoniae in children younger than 5 years: global es(mates. Lancet 2009;374:893-­‐902. For more informaPon contact: Hope Johnson ([email protected]) © Interna(onal Vaccine Access Center (IVAC) at the Johns Hopkins Bloomberg School of Public Health 2012 Al Rights Reserved