!005

~ these in

>gen-1989.

ty of.Iates915-c,:;,

,,;,

,,"'

)

IJIi!

'11

10

~A

10','-;

fils

ORIGINAL STUDIES .(5.J.J -:;.

RapidDiagnosisof PneumococcalMeningitisImplicationsforTreatmentand MeasuringDiseaseBurden

Samir K. Saha, PhD, *t Gary 1. Darmstadt, MD, t Noboru Yamanaka, MD, PhD,§Dewan S. BiIlal, MSc, *§ Tania Nasreen, MSc, * Maksuda Islam, EA, * and Davidson H Hamer, MDt

Background: Streptococcus pneumoniae is the leading cause ofchildhood pneumonia and meningitis worldwide. Isolation of thisorganism, however, is uncommon in resource-poor countries, in partbecause of extensive use of prior antibiotics. A rapid, highly sensi-tive immunochromatographic test (lCT) for S. pneumoniae wasevaluated for the diagnosis of meningitis.Methods: Cerebrospinal fluid (CSF) from 450 children with sus-pected meningitis was tested with ICT, and results were comparedwith CSF culture, latex agglutination test (LAT) and/or polymerasechain reaction (PCR). Serial CSF specimens from 11 patients werealso evaluated for duration of positive results during effectiveantimicrobial therapy.Findings: AU 122 cases of pyogenic pneumococcal meningitispositive either by culture (N = 87) or PCR (N = 35) were positiveby ICT, yielding 100% (122 of 122) sensitivity.AUpurulent CSFspecimens from patients with meningitis caused by other bacteria byculture (N = 149) or by LAT (N = 48) or those negative by culture,LAT and LytA and thus of unknown etiology (N = 20), and normalCSF specimens (N = 104) were negative by lCT. Thus the speci-

ficity onCT also was 100% (321 0021), although negativity onCTwas not confirmed by PCR, if it was positive for other organismseither by culture or LAT. Serotyping of S. pneumoniae strainsrevealed 28 different serotypes, indicating that outcome of ICT areindependent of diverse capsular serotype of pneumococcus, Antigenwas detected by ICT for at least 10 days after presentation, and 1was still positive on day 20, which was longer than for either LATor PCR.

Interpretation: ICT for pneumococcal antigen in CSF is 100%sensitive and specific in diagnosing pyogenic pneumococcal men-ingitis and can detect -30% more pneumococcal meningitis casesthan with culture alone. The simplicity of the test procedure and thelongevity of CSF antigen detection suggest the potential utility ofICT to estimate the true burden of pneumococcal disease, as for

Accepted for publication July 12, 2005.From the *Department of Microbiology, Bangladesh InstitUte of Child

Health, Dhaka Shishu (Children) Hospital, Dhaka, Bangladesh; thetDepartment of International Health, Bloomberg School of PublicHealth, Johns Hopkins University, Baltimore, MD; the tCenter forInternational Health and Development, Boston University School ofPublic Health Boston University, Boston, MA; and the §Department ofOtolaryngology, Wakayama Medica! University, Japan

Supported in part by Binax Inc., Portland, ME, and the U.S,-Japan Co-operative Medical Science Program, Japan,

Address for reprints: Samir K. Saha, PhD, Department of Microbiology, DhakaShishu (Children) Hospital, Bangladesh InstitUte of Child Health, Dhaka-1207, Bangladesh, Fax 8802 8611634; E-mail sksaha@hangla.net.

Copyright @ 2005 by Lippincott Williams & WilkinsISSN: 0891-3668/05/2412-1093DOl: 1O.I097/0I.inf.OOOOI90030,75892.78

Haemophilus i'!fiuenzae type b using data from meningitis, and toguide selection of appropriate antibiotic treatment, especially in re-source-poor countries with widespread prehospital antimicrobial use.

Key Words: immunochromatographic test, meningitis,pneumococcus, pneumonia, polymerase chain reaction,Streptococcus pneumoniae

(Pediatr Infect Dis J 2005;24: 1093-1098)

Pneumococcal pneumonia and meningitis are responsiblefor an estimated 800,000-1 million child deaths each

year.I-4 Meningitis frequently leads to chronic sequelae andincurs substantial direct and indirect costs.s Streptococcuspneumoniae as an etiology of pnewnonia largely goes unde-tected given that blood cultures lack sensitivity for detectingthe organism because most patients with pneumonia are notbacteremic. Pooled data from lung aspirate studies, per-formed mostly in developing countries, indicate that 55% ofinfections are caused by bacterial pathogens, predominantlyS. pneumoniae, Haemophilus injluenzae and Staphylococcusaureus.6 Lung puncture is not performed routinely, however,because of the invasive nature of the procedure and difficultyin obtaining ethical clearance. Therefore evaluation of pneu-mococcal meningitis cases might be useful to indirectlyestimate the overall burden of pneumococcal pneumonia,similar to the use of the rate of meningitis due to H. influen-zae type b (Hib) for estimating total Hib disease burden.7Isolation of pneumococcus, however, is uncommon in men-ingitis cases that occur in locations with limited resources andwhere the burden of disease is the greatest. This is partly theresult of an absence of standard microbiology procedures,which, in turn, makes clinicians reluctant to perform lumbar

punctures. Pneumococcal isolation is also jeofardized byprior antibiotic therapy which, in Bangladesh, is the rulerather than an exception. Thus the true burden of pneumo-coccal diseases is still unknown in many developing COull-tries,9 including Bangladesh.

Alternate methods to diagnose pneumococcal meningi-tis in cases with prior antibiotic use include detection ofantigen by latex agglutination test (LAT), or the LytA gene bypolymerase chain reaction (PCR). Antigen detection by LAT,however, is expensive, and the sensitivity varies dependingon the commercial assay. Moreover PCR, although consid-ered a standard, is still not widely available for rapid diag-nosis, especially in resource-poor settings.

ThePediatricInfectiousDiseasejournal. Volume 24, Number 12, December2005 1093

"

J.'Oi

'.''.:,"!

",J

.

Saha et at The PediatricInfectious Diseasejournal. Volume 24, Number 12, December 2005 The

In recent years, the immunochromatographic test (ICT)(Binax NOW Streptococcus pneumoniae test; Binax Inc.,Portland, ME) was developed with the aim to detect pneu-mococcal antigen in urine collected from patients with inva-sive pneumococcaldisease.to This rapid assay, unlike otherserotype-specific tests, detects the C-polysaccharide moleculefound in the cell wall of all pneumococci. 10.\\ A majorlimitation of this test, however, is that children in areas withhigh nasopharyngeal colonization rates often have positiveurinary antigen tests in the absence of any clinical manifes-tations of pneumonia or bacteremia.12~14In contrast, the sameassay used for detection of antigen in cerebrospinal fluid(CSF) was found to be 95~IOO% sensitive and 100% specificfor pneumococcal meningitis when compared with CSF cul-ture.15.J6 Although studies have demonstrated the potentialutility of this test for the detection of pneumococcal menin-gitis, major limitations of studies to date include small num-bers of culture-positivecases, mostly from adults, and dis-crepancies between lCT and culture-negative results were notvalidated (eg, with PCR).

The present study aimed to (1) determine whether ICTdetects pneumococcalantigen in urine from healthy childrenin Bangladesh, (2) investigate the utility of ICT of CSF forthe diagnosis of pneumococcal meningitis cases with the useof CSF culture and/or PCR as the standard, (3) elucidatewhether outcome of ICT varies with pneumococcal serotypeand (4) evaluate the duration of positive tests with LAT, ICTand PCR during effective antimicrobialtherapy.

.

METHODSFor study aim 1, urine specimenswere collected from

healthy children 2-9 months old who presented for routineimmunizations at Dhaka Shishu (Children) Hospital. Forstudy aims 2-4, patients younger than 5 years of age withclinical suspicionof meningitiswere enrolled.Patients found

to have progenic meningitis (CSF containing 2':100 leuko-cytes/mm with >50% neutrophils and/or growth of theorganism in culture) were subsequently followed. Anothergroup of patients with febrile convulsions and/or CSF with0-5 cells,negativeculture,normalprotein and glucose levelsand no evidence of meningitis on follow-up were includedas a control group. Equivocal cases with cell count from 6to 99/mm3and negative culture were excluded. Serial CSFspecimenswere obtainedfrom some patients as routinepractice of clinical consultantsto documenta decliningcell count before discharge or per protocol for a concurrent

World Health Orianization (WHO)-sponsoredclinical studyof meningitis. 17.1

Study procedures were approved by the BangladeshInstitute of Child Health Ethical Review Committee, andwritten informed consent was obtained from each patient'sguardian who accompaniedthe child to hospital.

Specimens of CSF were evaluated biochemically andcultured following standard procedures as described previ-ouslyP,19 LAT (Remel, Inc., Lenexa, KS) and ICT wereperformedaccordingto the manufacturers' instructions. ICTwas done on CSF samples by dipping a swab into the CSFspecimen,then insertingit into the test device. Reagent A, a

1094

buffer solution, was added from a dropper bottle, and thedevice was closed, bringing the sample into contact with thetest strip. Pneumococcalantigen present in positive samplesis captured by immobilized S. pneumoniae antibody, whichreacts to bind conjugatedantibody, forming the sample line.Immobilizedgoat anti-rabbit IgG also captures a visualizingconjugate, forming the control line. All test results (positiveand negative) are read visually at 15 minutes. The test wasinterpreted by the presence or absence of visually detectablepink to purple lines. A positive specimen produced both asample line and a control line, A negative result only pro-duced the control line, and the test was considered invalid ifthe control line did not appear.

LAT was performed on pyogenic CSF samples bymixing 40 ILLof boiled urine or CSF and I drop of test latexon the specifiedcircle of the supplied slide and then mixingon a rocking device for 3 minutes. LAT was done fordetection of antigens of S.pneumoniae, Hib, Neisseria men-ingitidisACY W135, groupB Streptococcusand Escherichiacoli Kl/N. meningitidis B. Two independent laboratory per-sonnel read the test results; S.K.S" who was blinded to otherlaboratory results, interpreted discrepant results. The LATwas interpreted as negative if it was clearly nonreactive orequivocal (:!:). To reduce costs, LAT was done on the firstday only if the Gram stain result was negative or equivocal.An aliquot of CSF was preserved, and LAT was performedon the second day if there was no growth from CSF.18

PCR to detect the LytA gene was performed in culture-negative, ICT-positive CSF specimens using the primers

described earlier: 5'-694TGAAGCGGA TTATCACTGGC71f3'; and 5'-966GCTAAACTCCCTG TATCAAGCG94s-3'. 0In brief, DNA was extracted from 200 ILL of CSF with theQIAmp kit (Qiagen, Hilden, Germany), and a 5-ILL aliquotwas placed in a PCR mixture containing 2.5 ILL of lOXreaction buffer, 2.0 ILL (2.5 mM) of each deoxynucleotidetriphosphate, 0.125 ILL (5 units/ILL) of Taq polymerase,0.025 ILLof LytAs, 0.025 ILLof LytAr (O.lILM) and 20.3 ILLof dH2O (Takara Bio, Inc" Shiga, Japan). DNA was dena-tured in a PX2 thermal cycler (Thermo Electron Corp.,Needham Heights, MA) at 94°C for 2 minutes followed by 30seconds at 94°C, 55°C and noc per cycle for 30 cycles.Further extension was done at noc for 10 minutes. Afteramplification, 10 ILL of PCR product were electrophoresedon a 3% agarose gel (Invitrogen, Carlsbad, CA). For inter-pretation, PCR product from CSF was compared with the273-bp LytA product of a previously characterized S. pneu-moniae isolate.

Pneumococcal strains were serotyped by the capsu-lar swelling procedure (quellung reaction) with antipneumo-coccal omni, pool, type or group and factor sera (StatensSeruminstitut, Copenhagen, Denmark) as described previ-ously?I,22 Nontypable S. pneumoniae strains were screenedout using omni sera at the first step of serotyping.

CSF specimens were tested for presence of antibioticfollowing the standard procedure described previously.23 Inbrief, a blank disc (Oxoid, Hampshire, United Kingdom) wasplaced on the lawn of a pansensitive organism (Micrococcusfetus ATCC 9341), and then the disc was soaked with 10 ILLof CSF. The plate was incubated overQight at 37°C, and any

zonthebra

waispe

. poscullismingby

~

FIC

cy1WE

I&:J2005 Lippincott Williams & Wilkins ~.

>Os

,<"~UW""\",,, ". f<.7,H'!i,D"""""""'-"'~XJE:';;~" :i:"£"" , ",",'

The Pediatric InfectiousDiseasejournal. Volume 24, Number 12, December 2005 Pneumococcal Meningitis

zone of inhibition around the disc was considered positive forthe presence of an aI1tibiotic capable of crossing the blood-brain barrier.

The standard for detennination of sensitivity of ICTwas positive culture or presence of LytA in culture-negativespecimens.24 For specificity, the comparison standard was

, positive culture for other organisms besides S. pneumoniae,culture-negative specimens positive by LAT for other organ-isms, normal CSF or specimens negative by all tests, includ-ing LyrA. Data were double entered, validated and analyzedby Epi Info 6.1.

RESULTS

Urine. Pneumococcal antigen was detected by ICT in urinespecimens from 51% (102 of 200) of healthy children.Cerebrospinal Fluid. Of 512 patients younger than 5 yearsold evaluated by lumbar puncture for clinical suspicion ofmeningitis at Dhaka Shishu Hospital from 2001 to 2004,346(68%) were enrolled with features of pyogenic meningitis; 62cases had equivocal features of meningitis and were ex-cluded; in 104 cases, meningitis was ruled out, and these wereincluded as controls (Fig. I). The mean and median ages of

FIGURE1. Studyprofiledemonstrating resultsof CSFtesting stratifiedby baselineCSFcellcounts. TLCindicatestotal leuko.cyte count; Cult, culture; Neg, negative; Pos, positive; SPN, S. pneumoniae; Nmen, N. menlngitidis; +ve, positive; *, 7 cultureswere not available for further tests.

(Q2005 Lippincott Williams & Wilkins 1095

thetheliesich.ne.ingivevasbletI aro-i if

bytexingforen-ia

er-tIerL\Tor

Irst:al.led

II:)ers

IlO

'.

.*,.

lA-w

.

Cd:r-Ilev-

I-! u-.0-DBti-ed

cInas""

'Y

'"

CerebrospinalFluid(N=512)

1TLC 06 -99/mm3 CultNeg

Pyogenic (N= 346) I

'

I Normal (N= 104) IN=62

.., ICTNegCuitNeg

1104/104; 100%

Growth Prior Antibiotic +ve No Growth H Prior Antibiotic+ve

(N= 236) N= 22 (9%) (N= 110*) N=cS9(8)%)

-II H. influenzae .. IcrNeg LArHiblcrNeg(N= 125) 125/l25; 100% (N= 41)

LA rspnI CTPoS LytAPOSS. pneumoniae f-t

-I (N= 87)ICrPos (N= 30) 30/30; 100%87/87; 100%

1+LArNmelllcrNeg

N. meningitidis IcrNeg(N= 7)

(N= 6) -II 6/6; 100%LA rNegICTPos LytAPOS(N= 5) 5/5; 100%

Others (N= 18) ICTNeg-II 18/18; 100%

1+ LA TNegICTNegI---iI

LytANeg

(N= 20) 20120; 100%

Saha et at The PediatricInfectious Diseasejournal. Volume24, Number 12, December2005 Th.

.

the 450 enrolled cases were 9 and 7 months, respectively,with 90% of the cases in the first year of life. Agedistribution was similar in pyogenic cases and in caseswith normal CSP. There was a marked predominance ofboys [59% (302 of 512)].

Figure 1 also shows the results of CSF specimensevaluated by culture, ICT, LAT and PCR for LytA. Among346 pyogenic meningitiscases, 236 (68%) yielded growth inculture. Presence of antibiotic was detected in 32% (Ill of346) of pyogenic CSF specimens, including 9% (22 of 236)and 8I% (89 of 110)of culture-positiveand culture-negativegroups, respectively.Among culture-positivecases, antibiot-ics had been used before lumbarpuncture in no instance(0 of87) of infectionwith S. pneumoniae, whereas in 14% (18 of125) of cases of H. injluenzaemeningitis,prior antibiotic usewas detected.

Overall 122 (35% of 346) cases of pneumococcalmeningitis were identified, including culture-positive cases(N = 87) plus culture-negative cases in which PCR waspositive, that is, detection of LytA (N = 35).Thussensitivityof culture for identifYingS. pneumoniae was 71% (87 of122). All cases in which S. pneumoniae was recovered byculture ofCSF (N = 87) or identified by LytA (N = 35) werepositive by ICT within :'52 minutes. The sensitivity of ICTtherefore was 100% (122 of 122) for detection of S. pneu-moniae. In contrast, all CSF specimens were negative by ICTfrom meningitis cases that were culture-positive for otherorganisms (N = 149), which showed no growth in culture butwere identifiedas H. injluenzae(N = 41) or N meningitidis(N = 7) by LAT of CSF, or were normal (ie, had normalcytology, biochemistry and negative culture) (N = 104). Forthese cases, however, LytA was not done to confirm negativ-ity for pneumococcal genome. In addition, all specimens thatwere negative by all tests, including LytA (N = 20), were alsonegative by ICT. Thus the specificity of ICT under theseconditions was also 100% (381 of 381).

Among 35 culture-negative pyogenic CSF samplespositive by JCT and LytA for S. pneumoniae, LAT waspositive in.30 but falsely negative in 5 cases.

Among 87 pneumococcal isolates from CSF culture, 63were available for serotyping. Testing revealed 28 differentserotypes with a relative predominance of 2 (N = 7), 5 (N =7),33 (N = 5) and nontypable (N = 6) strains. Because LATwas conducted selectively, based on first day Gram stainresults, only 41 pneumococcal culture-positive cases had asimultaneous assay for LAT and, in contrast to ICT, whichwas universally positive in these cases, 6 (15%; 6 of 41) CSFspecimens did not show any agglutination. Among thesefalsely LAT-negative cases, 3 were nontypable whereas theothers were types 1, 12A and 20.

Markers for S. pneumoniae were followed in serial CSFspecimens from 11 patients from whom 2::3 specimens wereavailable with sufficient quantity to run the assays, LAT, PCRand ICT, in addition to cytology and biochemistry. All werepositive by ICT for at least 10 days after the initiation ofappropriate treatment, and 1 patient was still positive on day20 (the specimen available with longest duration) (Table 1).All patients survived on first line antibiotic therapy. LytA

1096

r--

gene was undetectable by PCR by day 19-20 of treatment,whereas LAT was negative as early as day 10.

T'"-

DISCUSSIONS. pneumoniae and Hib are the predominant causes of

meningitisand pneumonia in Bangladesh.17,18Recent studiesshowed an increasing trend of nonsusceptibility of H. injlu-enzae strains.18On the other hand, S. pneumoniae strains

have remained susceptible to lenicillins, and most otherdrugs for more than 1 decade.I ,19,25Therefore rational em-piric treatment of these 2 organisms differs; thus rapid etio-logic diagnosis would be of great interest to rationalizeempiric therapy.

Our detection of S. pneumoniae antigen in urinespecimensfrom one-half (51%)of healthy children by lCT issimilar to the finding of others12-14and is most likely a re-sult of the high pneumococ(.(alnasopharyngeal colonizationrates25 and high prevalence of pneumonia in our popula-tion.26Therefore antigen detection in urine by ICT is notuseful for predicting the presence of disease in Bangladeshichildren.

Jn contrast, this study showed that the JCT test was100% sensitive fo~identification of S. pneumoniae in pyo-genic CSF with ve"ty,high specificity compared with thenonpneumococcalcases proved by cultureand/or LAT and/orLytA. The findings of this study confirm the results of thepreviousstudieswith a small numberof meningitiscases.15,16Our data demonstrate that CSF remained ICT-positive evenwhen the organism, which had originally been recoveredfrom culture, was eradicated and CSF become culture-nega-tive (Table 1). This finding was further strengthened by theinvestigations with serial CSF specimens from 11 culture-positive cases. PersistentOdetectionof S.pneumoniaeantigenby lCT for at least as long as the presence of LytA substan-tiates the high sensitivity of ICT in detecting pneumococcalC-polysaccharideantigen while concurrently demonstratingthe longevityof antigen detection in CSF from children withmeningitis. There has not been any previous research on theduration of antigen presence in CSF samples in subjects withpneumococcalmeningitis. Antigen detection in urine by JCTremains positive for as long as 89 days in adults withcommunity-acquiredpneumococcal pneumonia,z7,28This ispossibly causedby the releaseof small fragmentedmoleculesof C-polysaccharide, such as teichoic acid, at the time of

antibiotic theraJ'Y'which escape the phagocytic mechanismsof the host.29.3Nevertheless the high false positivity rate ofICT in urine tests obviates its utility for pneumonia casedetection in adults who have had recent episodes of lowerrespiratory tract infections. False positive urinary antigentests can occur with even greater frequency in children as aresult of nasopharyngealcolonizationand/or recent pneumo-coccal respiratory tract infections.

In this series of patients, prior antibiotictherapy, basedon the presence of antibiotic in CSF specimens at sufficientlevels to inhibit bacterial growth, was found in one-third ofpatients with pyogenic meningitis. Detection of prior antibi-otic use in no cases of pneumococcal-meningitis identifiedby culture (0 of 87) but in 14% (18 of 125) of cases in

'!~'

,~..

ej.

";,,

wbonlstr~carpncsuene~

utadit1witanewe]the

casforimeithi~the:sonthedet,thetha'

~ 2005 Lippincott Williams& Wilkins rQ2

)05

mt,

".IJ

ThePediatricInfectiousDiseasejournal. Volume24, Number 12, December2005. --- ---" Pneumococcal MeningJtis

Patient Characteristics

TABLE 1. Test Resnlts of Serial CSF Specimens From 11 Patients

Sample Characteristics

~

...~'-i

e8df'lisf#'se'CI'mk

rf.

=dntofIi-~din

,.,

which Hib was cultured from CSF suggests that relianceonly on culture positivity in this population, in which moststrains of S. pneumoniae are penicillin-susceptible,17,19,25can specificallylead to an underestimationof the burden ofpneumococcal disease.31The ICT, however, can be used tosuccessfullydetectpneumococcalmeningitis in these culture-negative cases.

Pneumococcal serotypes among the Bangladeshi pop-ulation are diverse.21.25Nevertheless, in this series, all 28different serotypes of 25 serogroups gave positive resultswith ICT. In contrast, 15%(6 of 41) of culture-positivecasesand 14% (5 of 35) of culture-negative,LytA-positive caseswere negative by LAT, thus implying a lower sensitivity ofthe later test.

Our study has a few limitations. (1) Culture-negativecases with 6-99 cells/mm3were not subjectedto ICT; there-fore the utility of this test for detection of pneumococcalmeningitis in patients with CSF white blood cell counts inthis range (6-99/mm3) could not be discerned. Although allthese cases were negative by culture, there might have beensome pneumococcalcases as indirectlyevidenced for Hib inthe recent Lombock vaccine tria1.32(2) We also could notdetermine the sensitivity and specificity of LAT, althoughthe available data suggest that its sensitivity is lower thanthat of ICT.

@ 2005 Lippincott Williams & Wilkins

This study evaluated ICT with large number of casesusing detection of LytA gene as a standard to confirm culture-negative and ICT-positive cases of pneumococcal meningitisand supported the preliminary findings of Samra et all6 andMarcos et al.ls Although PCR for the pneumococcal genewas not applied to culture-negative, LAT-positive nonpneu-mococcal cases, ICT proved to be similar to detection of LytAby PCR for diagnosis of pneumococcal meningitis. PCR,however, requires advanced facilities and expertise and in-curs higher costs. On the other hand, ICT can be done at thebedside or field level without any laboratory facility, and thedevice can be stored at room temperature. Given the simplic-ity of the test and the rapidity with which results becomeavailable, this test can be used at hospitals of all levels, fromprimary to tertiary, for surveillance on the burden of pneu-mococcal meningitis. This information in turn could be usedto estimate the burden of pneumococcal pneumonia as hasbeen done for H. influenzae.7 In addition to disease burden,ICT might have paramount significance in determining thetreatment of pneumococcal meningitis in Bangladesh andneighboring countries. where prior empiric treatment is com-mon and isolation of S. pneumoniae is jeopardized even aftera single dose of first line antibiotic because of its highsusceptibility to penicillin.2S,33

1097

Patient! Antibiotic Outcomeof Hospital Culture LAT PCR for Lyt A ICTSample Given Treatment Day: of 1 Ceftriaxone Sequelae 0 Streptococcus pneumoniae + + +lies 4 No growth + + +flu. 10 No growth + + +ins 20 No growth - - +

2 Ceftriaxone Cured 0 S. pneumoniae + + +her 4 No growth + + +:tn- 10 No growth + + +

10-3 Ampicillinand Cured 0 S. pneumoniae + + +

chloramphenicol 6 No growth + + +ize 12 No growth - + +

4 Ceftriaxone Cured 0 S. pneumoniae + + +

ine10 No growth + + +17 No growth - + +

is 5 Ceftriaxone Cured 0 S. pneumoniae + + +re- 4 No growth + + +

10 No growth + + +lon 6 Ampicillinand Sequelae 0 S. pneumoniae + + +.1a- chloramphenicol 4 No growth + + +

10 No growth + + +110t 19 No growth - - +shi 7 Ceftriaxone Cured 0 S. pneumoniae + + +

4 Nogrowth + + +

I10 Nogrowth - + +8 Ceftriaxone Sequelae 0 S. pneumoniae + + +

6 No growth + + +12 No growth + + +

9 Ceftriaxone Cured 0 S. pneumoniae + + +4 No growth + + +

tile 10 No growth + + +tN 16 No growth . + +

i10 Ceftriaxone Cured 0 S. pneumoniae + + +4 No growth + + +

10 No growth + + +11 Ceftriaxone Cured 0 S. pneumoniae + + +5 No growth + + +

\12 No growth + + +

/jI

"f,:en

Saha et al ThePediatricInfectiousDiseasejournal. Volume24, Number 12, December2005

I

IREFERENCES

1. World Health Organization. Pneumococcal vaccines. Wkly EpidemiolRecord. 2003;14:110-119.

2. MuIlholiand K. Strategies for the control of pneumococcal diseases.Vaccine. 1999;100I:S79-S84.

3. Williams BG, Gouws E, Boschi-Pinto C, Bryce J, Dye C. Estimates ofworld-wide distribution of child deaths from acute respiratory infections.Lancet Infect Dis. 2002;2:25-32.

4. Greenwood B. The epidemiology of pneumococcal infection in childrenin the developing world. Phil Trans R Sac Lond B. 1999;354:777-785.

5. Baraff U, Lee ST, Schrieger DL Outcomes of bacterial meningitis inchildren: a metaanalysis. Pediatr Infect Dis J. 1993;12:389-394.

6. Vuori-Holopainen E, Peltola H. Reappraisal oflung tap: review of an oldmethod for better etiologic diagnosis of childhood pneumonia. CUnInfect Dis. 2001;32:715-726.

7. World Health Organization. Expert Review of a Tool for RapidlyAssessing Hib Disease Burden. Geneva, Switzerland: WHO Headquar-ters; 2000.

8. Banish M. The Bangladesh drug policy: the next step: using good drugsgoodly. Bangladesh J Child Health. 1992;11:63-72.

9. Greenwood. BM. The epidemiology of acute bacterial meningitis intropical Afiica. In: Bacterial Meningitis. London, United Kingdom:Academic Press; 1987:61-87.

10. Dominguez J, Gali N, Blanco S, et al. Detection of Streptococcuspneumoniae antigen by a rapid immunochromatographic assay in urine.Chest. 2001;119:243-249.

11. Gillespie SH, Smith MD, Dickens A, Raynes JG, McAdam KPWJ.Diagnosis of Streptococcus pneumoniae pneumonia by quantitativeenzyme-linked immunosorbent assay of C-polysaccharide antigen.J Glin Pathol. 1994;47:749-751.

12. Adegbola RA, Obara SK, Biney E, Grrenwood BM. Evaluation of BinaxNOW Streptococcus pneumoniae urinary antigen test in children in acommunity with a high carriage rate of pneumococcus. Pediatr InfectDis J. 2001;20:718-719.

13. DoweIl ST, Garman RL, Liu Gang, Levine OS, Yang YH. Evaluation ofBinax NOW, as assay for the detection of pneumococcal antigen in urinesamples performed among pediatric patients. Clin infect Dis. 2001;32:824-825.

14. Hamer DH, Egas J, Estrella B, MacLeod WB, Griffiths JK, SemperteguiF. Assessment of Binax NOW Streptococcus pneumoniae urinary anti-gen test in nasopharyngeal pneumococcal carriage. CUn Infect Dis.2002;34: 1025-1028.

15. Marcos MA, Martinez E, Almela M, Mensa I, ADta MTJ. New rapidantigen test for diagnosis of pneumococcal meningitis. Lancet. 2001;357:1499-1500.

16. Samra Z, Shmuely H, Nahum E, Paghis 0, Ben-Ari 1. Use of the NOWStreptococcus pneumoniae urinary antigen test in cerebrospinal fluid forrapid diagnosis of pneumococcal meningitis. Diag Microbiol Infect Dis.2003;45:237-240.

17. Saha SK, Rikitomi N, Ruhulamin M, et al. Antimicrobial resistance and

serotype distribution of Streptococc'us pneumoniae strains causing child-hood infections in Bangladesh, 1993-1997. J CUn Microbiol. 1999;37:798-800.

w

1098

18. Saba SK, Baqui AH, Darmstadt GL, et aI. Invasive Haemophilusinfluenzae type b diseases in Bangladesh, with increased resistance toantibiotics. J Pediatr. 2005;146:227-233.

19. Saba SK, Khan WA, Hoq MS, et al. Penicillin resistant pneumococci inBangladeshi children. Lancet. 1991;337:734-735.

20. Ubukata K, Ashi Y, Yamane A, Konno M. Combinational detection ofautolysin and penicillin binding protein 2B genes of Streptococcuspneumoniae by PCR. J CUn Microbiol. 1996;34:592-596.

21. Saha SK, Rikitomi N, Biswas D, et al. Serotypes of Streptococcuspneumoniae causing invasive childhood infections in Bangladesh, 1992to 1995. J CUn Microhiol. 1997;35:785-787.

22. Facklam RR, Washington JA 11. 1991. Streptococcus and related cata-lase-negative Gram positive cocci. In: Balows A, Hausler WJ Jr, Her-mann K, Isenberg HD, Shadomy HI, eds. Manual of Clinical Microbi-ology. 5th ed. Washington, DC: American Society for Microbiology;238-257.

23. Markowitz M, Gordis L. A mail-in technique for detecting penicillin inurine: application to the study of maintenance of prophylaxis in rheu-matic fever patients. Pediatrics. 1968;41: 151-153.

24. Messmer TO, Sampson 1S, Stinson A, Wong B, Carlone GM, FacklamRR. Comparison off our polymerase chain reaction assays for specificityin the identification of Streptococcus pneumoniae. Diagn MicrobiolInfect Dis. 2004;49:249-254.

25. Saha SK, Baqui AH, Darmstadt GL, et al. Comparison of antibioticresistance and serotype composition of carriage and invasive pneumo-cocci among Bangladeshi children: implications for treatment policy andvaccine formulation. J CUn Microhiol. 2003;41:5582-5587.

26. Rudan I, Tomaskovic L, Boschi-Pinto C, Campbell H, WHO HealthEpidemiology Group. Global estimate of the incidence of clinical pneu-monia among children under five years of age. Bull World HealthOrgan. 2004;82:895-903.

27. Ishida T, Hashimoto T, Arita M, Tojo Y, Tachibana H, Jinnai M. A3-year prospective study of a urinary antigen-detection test for Strepto-coccus pneumoniae in community-acquired pneumonia: utility and clin-ical impact on the reported etiology. J Inject Ghemother. 2004;10:359-363.

28. Murdoch DR, Laing RTR, Cook 1M. The NOW S. pneumoniae urinaryantigen test positivity rate 6 weeks after pneumonia onset and amongpatients with COPD. Clin Infect Dis. 2003;37:153-154.

29. Fischer H, Tomasz A. Production and release of peptidoglycan and wallteichoic acid polymers in pneumococci treated with 13-lactam antibiotics.J Bacteriol. 1984;157:507-513.

30. Stuertz K, Merx I, Eiffert H, Schmutzhard E, Mader M, Nau R. Enzymeimmunoassay detecting teichoic and lipoteichoic acids versus cerebro-spinal fluid culture and latex agglutination for diagnosis of Streptococ-cus pneumoniae meningitis. J Clin Microbiol. 1998;36:2346-2348.

31. ShanD F. Bacterial pneumonia: commoner than perceived. Lancet. 2001;357:2070-2072.

32. Gessner BD, Sutanto A, Linehan M, et al. Incidences of vaccine-preventable Haemophilus infuenzae type b pneumonia and meningitis inIndonesian children: hamlet-randomised vaccine-probe trial. Lancet.2005;365:43-52.

33. Invasive Bacterial Infection SurveilIance (IBIS) Group. Prospectivemulticentre hospital surveillance of Streptococcus pneumoniae diseasein India. Lancet. 1999;353:1216-1221.

t"

...:!!

Ba.mo:ranR.j(NIMefacbacturtinvRepicha(casfaeprein~intIUS(ide:rerweC~m(th(

K.int

(Po

.1baGpi.

~

Ie.ccISDC

@ 2005 Lippincott Williams & Wilkins 11

-