w w w . e l s e v ie r . c o m / l o c a t e / b j i d
The
Brazilian
Journal
of
INFECTIOUS
DISEASES
Original
article
Serum
markers
as
an
aid
in
the
diagnosis
of
pulmonary
fungal
infections
in
AIDS
patients
Ana
Isabela
Morsch
Passos
a,
Rachel
Polo
Dertkigil
b,
Marcelo
de
Carvalho
Ramos
a,
Ariane
Fidelis
Busso-Lopes
d,
Cibele
Tararan
a,
Erivan
Olinda
Ribeiro
a,
Angélica
Zaninelli
Schreiber
c,
Plinio
Trabasso
a,
Mariangela
Ribeiro
Resende
a,
Maria
Luiza
Moretti
a,∗aUniversidadedeCampinas,FaculdadedeCiênciasMédicas,DepartamentodeMedicinaInterna,Campinas,SP,Brazil bUniversidadedeCampinas,FaculdadedeCiênciasMédicas,DepartamentodeRadiologia,Campinas,SP,Brazil cUniversidadedeCampinas,FaculdadedeCiênciasMédicas,DepartamentodePatologiaClínica,Campinas,SP,Brazil dLaboratórioNacionaldeBiociências(LNBio),Campinas,SP,Brazil
a
r
t
i
c
l
e
i
n
f
o
Articlehistory: Received9March2017 Accepted9July2017 Availableonline29July2017
Keywords: Pulmonaryinfection HIV/AIDS (1-3)--D-Glugan LDH LAMP Pneumocystosis Fungalinfection
a
b
s
t
r
a
c
t
Introduction:TheetiologyofpulmonaryinfectionsinHIVpatientsisdeterminedbyseveral variablesincludinggeographicregionandavailabilityofantiretroviraltherapy.
Materialsandmethods: Across-sectionalprospectivestudywasconductedfrom2012to2016 toevaluatetheoccurrenceofpulmonaryfungalinfectioninHIV-patientshospitalizeddueto pulmonaryinfections.Patients’serumsweretestedfor(1–3)--D-Glugan,galactomannan, andlactatedehydrogenase.Theassociationamongthevariableswasanalyzedbyunivariate andmultivariateregressionanalysis.
Results:60 patients were included in the study. The patients were classified in three groups:Pneumocystisjirovecii pneumonia (19patients), community-acquired pneumonia (18patients),andotherinfections(23patients).Theoverallmortalitywas13.3%.Thetime sincediagnosisof HIVinfection wasshorterin thepneumocystosis group (4.94 years; p=0.001)than forthe other twogroups ofpatients. The multivariateanalysis showed thathigher(1-3)--D-Glucanlevel(mean:241pg/mL)andlactatedehydrogenase (mean: 762U/L)wereassociatedwiththediagnosisofpneumocystosis.Pneumocystosiswasthe aids-definingillnessin11outof16newlydiagnosedHIV-infectedpatients.
Conclusion:Intheeraofantiretroviraltherapy,PJPwasstillthemostprevalentpulmonary infectionand(1-3)--D-Glucanandlactatedehydrogenasemaybesuitablemarkerstohelp diagnosingpneumocystosisinourHIVpopulation.
©2017SociedadeBrasileiradeInfectologia.PublishedbyElsevierEditoraLtda.Thisisan openaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/ by-nc-nd/4.0/).
∗ Correspondingauthor.
E-mailaddresses:moretti.luiza@gmail.com,mlmoretti@hc.unicamp.br(M.L.Moretti). http://dx.doi.org/10.1016/j.bjid.2017.07.002
1413-8670/©2017SociedadeBrasileiradeInfectologia.PublishedbyElsevierEditoraLtda.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Introduction
The epidemiology of pulmonary infections varies
accord-ing to geographicalregions and to whether the infection
occurredbeforeorafteractiveantiretroviraltherapy(ART)
availability. In some low-income countries, tuberculosis
andpneumocystosisarestillthemostfrequentpulmonary
complicationsinHIV/AIDSpatients.1Conversely,inhigher
resources settings, the frequency of pulmonary diseases
associatedwithAIDShasdecreasedduetothebenefitsofthe
ART.2–4Hospitalizationduetochronicobstructiverespiratory
disease,lungcancer,andbacterialpneumoniahasbecome
more frequentthan pulmonaryinfections due to
Pneumo-cystisjirovecii,tuberculosis,andcytomegalovirus.5,6However,
lowerrespiratorytractinfections(LRTI)arestill25-foldmore
commonin the HIV populationcompared to the general
communitycausinganestimatenumberof20–25episodes
per100hospitalizationsworldwide.6,7
FungalinfectionsinHIV-infectedpatientsareneglected
diseases,predominantlyincountrieswithlimitedresources,
andrepresentasignificantcauseofpulmonaryinfections.
TheburdenofHIV-relatedmycosisworldwideisestimated
toaccount,peryear,formorethan950,000casesof
crypto-coccosis,400,000casesofPJP,and300,000ofdisseminated
histoplasmosis.8
Themajorityofthediagnosesofpulmonarydiseasesare
basedonclinicalsymptomsandX-rayfindingsresultingin
patientsbeingtreatedempirically.Identificationofthe
eti-ologicagentisa complex task requiring thecombination
ofmicrobiologicexams,serologicmarkers,molecular
tech-niques,andinvasiveprocedures, such aslungbiopsyand
bronchoalveolar lavage (BAL). New molecular techniques
havebeenstudiedtoimprovethediagnosisofpulmonary
infiltrates in the HIV-infected population. Loop-mediated
isothermal amplification (LAMP) has been evaluated for the
detection of Pneumocystis DNA in respiratory specimens
and could serve as a diagnostic tool.9 Serum (1-3)-
-D-Glugan(BG)hasbeenapromisingnon-culturemethodfor
the diagnosis of some fungal infections, including,
Can-didaspp.,Fusarium,andP.jirovecii.Recentdatahaveshown
that serum BG could be correlated with the diagnosis of
PJP10–12withasensitivityrangeof90–100%andspecificityof
65–100%.13–15
Inlow/middleincomecountriesthereisanurgentneed
for prospective studies to clarify the infectiouscauses of
pulmonary diseasesthat lead to hospitalizations in AIDS
patients.In this scenario,the properidentification of the
causativeagentswillindicateabettertherapeuticapproach
andimprovetheunderstandingoftheepidemiologyof
pul-monary affections responsible for hospitalizations in our
country.
Theaimofthisstudywastoevaluatetheoccurrenceof
pulmonaryfungalinfectionsandtheroleofserummarkers
inHIV-infectedpatientshospitalizedwithacuterespiratory
symptoms,inatertiarycarereferralhospitalinCampinas,
SaoPaulo,Brazil.
Materials
and
methods
Studypopulation
Weconductedaprospectivecross-sectionalstudy,from2012 to2016,attheHospitaldasClinicasoftheUniversityof Cam-pinas (UNICAMP), Sao Paulo,Brazil, whichis the reference hospitalformorethansixmillioninhabitants.
TheEthicalCommitteeapprovedthisstudy(No.8876/2012) andallpatientssignedtheinformedconsentform.The inclu-sioncriteriaincludedHIVinfection,ageover18yearsold,and hospitalization due tosymptoms oflower respiratory tract infection. Exclusion criteria includedpregnant women and patientswithnosocomialpulmonaryinfections.Atthetime ofhospitalization,aphysiotherapist(AIPM)performeda pul-monaryexaminationinallpatientsandcollectedsputumand orallavageforthemoleculardiagnosisofP.jirovecii.A radi-ologist assessed all chest radiographies and tomographies. For theassessmentofthepneumoniaseverity,thepatients wereclassifiedusingCURB-65score.16Atthedayofadmission,
thefollowingdatawereabstractedfromthepatients’records: age, gender, duration of HIV infection (in years), duration of hospitalization (in days), outcome,previous opportunis-ticinfections,hemoglobinlevel,hematocrit,totalleukocytes, serumcreatinine,serumurea,CD4Tcellcount,andHIVviral loadinthelasttwomonths.Inaddition,serumcryptococcal antigen, serologyforparacoccidioidomycosis, blood culture forbacteriaandfungi,cultureformycobacteriaandfungiin sputum,serumLDH,BGandGM,LAMPofrespiratory spec-imens(sputum,BAL)andorallavage(OL)forP.jiroveciiwere alsotested.
TheclinicaldiagnosisofpneumoniaduetoP.jiroveciiwas establishedaccordingtoWHOclinicalcriteriaofcase defini-tionforHIV-relatedopportunisticdiseases,17andachestX-ray
showingbilateral interstitialinfiltratesor achest tomogra-physhowingalterationscompatiblewithPJP(bilateralpatchy groundgrassopacitywithacentralperihilarpredominance). A definitive diagnosis was based on the identification of P. jirovecii incytology or immunofluorescent microscopy of inducedsputumorBALorhistologyoflungtissue.
Community-acquiredpneumonia(CAP)wasdefinedasthe presenceofcoughtogetherwithoneormoreofthesymptoms: chestpain,dyspnea,presenceofnewpulmonaryinfiltrateon chestradiography.18CriteriafordiagnosisofCAPandLRTIdue
tobacteria,fungiotherthanP.jirovecii,orparasiteswerebased ontheidentificationoftheetiologicagentinculturesofblood andrespiratorysecretions.
BGandGMinserumandinBALspecimen
Serumand BAL sampleswere tested forGMand BG. Con-centrations of GM were determined by using the Platelia Aspergillus Ag assay (Bio-Rad, Marnes-la-Coquette, France), accordingtothemanufacturer’sinstructions.AnOD≥0.5was consideredasapositiveresult.BGlevelsweredeterminedby theFungitell®assay(AssociatesofCapeCod,Inc.,CapeCod, MA,USA),accordingtothemanufacturer’srecommendations. BGlevels≥80pg/mLwereconsideredaspositiveresults.
60 patients included
Serum markers Imaging exams LAMP for PJ Cultures
Serum beta glucan (n=55) Serum Galactomannan (n=56) Serology for paracoccidioidomycosis (n=40) Cryptococcal antigen (n=42) Chest radiography (n=60) Chest tomography (n=34) In blood (n=37) In sputum for Mycobacterium (n=42) In sputum for In sputum for fungi (n=45) fungi (n=44) 67 patients screened Patients excluded 1 endocarditis
4 hospital acquired pneumonia 2 refused to participate
Serologies Direct microbiology
LDH (n=55) In sputum (n=21) In oral lavage (n=19) In BAL (n=2) In sputum for Mycobacterium (n=48)
Fig.1–Patientdispositionbyaccordingtoresultsoftheexamsperformed.
LAMPforP.jirovecii
Sputum,orallavage,andBALsamplesweretestedforP.jirovecii by LAMP. DNA extraction procedures strictly followed the instructionsdefinedbythemanufacturer(EikenChemicalCo., Tokyo, Japan)using theoligonucleotide primersspecificfor P.jirovecii.19
Statisticalanalysis
Categoricalvariableswereanalyzedusingchi-square,Fisher, and Mid-P exact tests, and for continuous variables t-test, ANOVA and Mann–Whitney/Wilcoxon test were used. A p-value<0.05wasconsideredstatisticallysignificant.The mul-tivariatestepwiselogisticregressionanalysiswasperformed todeterminethefactorsindependentlyassociatedwithfungal infection.Thefollowingprogramswereused:Epiinfov.7 (Cen-tersforDiseaseControlandPrevention,Atlanta,USA),Open SourceEpidemiologicStatisticsforPublicHealth,version3.03a updated2015/05/04,20andSASSystemforWindows(version
9.1.3,SASInstitute,Cary,NorthCarolina).
Results
Atotalof67HIV-infectedpatientswereincluded.Fig.1shows thedistributionofpatientsaccordingtoresultsoftheexams performed(imaging,biomarkers,LAMP,serologies,and micro-biologicexams).
Thepatientswereclassifiedintothreegroups:(a)PJP(19 cases):includedpatientswithPJPandpatientswithdiagnosis ofPJPplusanotherpulmonaryinfection/disease;(b)patients withCAP(18cases);and(c)otherdiagnosis(23casesoflower respiratoryinfection:ninecasesoftuberculosis,fivecasesof histoplasmosis,twocasesofnocardiosis,onecaseof cryp-tococcosis, one case of disseminated strongyloidiasis, one caseofpulmonaryembolism,andthreeundiagnosedcases). PJPandCAPwere themostfrequentdiagnosisrepresenting 31.6%and 30%,respectively. Tuberculosiswas diagnosedin five (8.3%) patients. Seven patients with CAP had positive
blood culturesforbacteria (one Acinetobacterbaumanii, one Kokuria micrococcus,andfive Streptococcuspneumoniae). Sero-logyforparacoccidioidomycosisanddetectionofcryptococcal antigeninserumwereassessedin40and42patients, respec-tively.Onlythepatient,withpulmonarycryptococcosis,had a positive cryptococcal antigen. All serologies for paracoc-cidioidomycosisturnedoutnegative.Thetwopatientswith histoplasmosishadculturesofskinbiopsiespositivefor Histo-plasma.
Directsputumsmearmicroscopyandsputumculturefor mycobacteria wereexamined in48and 42patients respec-tively,andforfungiin44and45patients,respectively.Direct sputumsmearwaspositiveforfungiinsevencases(fiveyeasts and filamentous fungi,one Strongyloides stercoralis and one non-pathogenicyeast)andforMycobacteriuminfourpatients. Sputumcultureforfungiwaspositivein24 cases(18 non-pathogenicyeast,threeCandidaalbicans,oneActinomycesspp., oneCryptococcusneoformansandoneHistoplasmacapsulatum) andforMycobacteriuminfourcases(twoMycobacterium tuber-culosisandtwoMycobacteriumavium).
The mean values of hemoglobin and hematocrit were below the referencevalues (Table1). Forty-onepercent(25 patients) werefemale. Themedian timesinceHIV diagno-sis was 10 years; however, 22.8% of the patients had less than oneyear. HIV infection was first diagnosedat hospi-talization in16 (26.6%)patients; PJP wasthe AIDS-defining illnessin11patients;themeanageofpatientswas43.4years (female:43.8;male:43.1);42(70%)patientshadlessthan250 cell/mm3CD4cellcountandonly14(23.3%)patientshadHIV viral load <50copies/mL.Themeans ofHIVviral load and CD4Tcellcountformales(760,261copies/mL;214cells/mm3) and females (291,703 copies/mL; 174 cells/mm3) were not significantly different (p=0.43 and p=0.56, respectively) (Table1).
Meantimeofhospitalizationwas19days.The length-of-stay inhospitalwas notassociated withoutcome (p=0.21; death:27.9days;survival:16.3days).Nine(15%)patientsdied duringhospitalization(4PJP,3CAP,1tuberculosis,and1 dis-seminatedstrongyloidiasis).Thefollowingvariableswerenot associatedwiththeoutcome:age(p=0.40),gender(p=0.46),
Table1–Laboratorydataof60HIV-infectedpatientshospitalizedwithpulmonaryinfections.
Exams Mean(±standarddeviation) p-Value Median
Female(n=25) Male(n=35) Female Male
Leukocytes,cells/mL 8144(±5578) 10,144(±11,402) 6600 7210 Hemoglobin,g/dL 9.66(±2) 11.1(±6.06) 9.8 11.3 Hematocrit,% 29.54(±5.9) 33.3(±6.88) 28.9 35.7 Platelets,cells/mL 237,920(±111,476) 244,428(±134,035) 237,000 213,000 Creatinine,mg/dL 1.49(±2.45) 1.8(4.4) 0.78 0.94 Urea,mg/dL 47.25(±58) 49.6(±39) 32.5 39 Lactatedehydrogenase(LDH)U/L 457(±207) 589(±392) 0.39 440 442
Viralload,copies/mL 291,703(±591,325) 760,261(±2,139,278) 0.43 52,804 72,391
CD4+lymphocytes,cell/mm3 174(±209) 214(±297) 0.56 139 64
Referencevalues.Creatinine:adults–men<1.20mg/dL; women<0.90mg/dL; urea:adult≤65years: 16.6–48.5mg/dLandadult>65years: <71mg/dL;LDH:adult≤65years:240–480U/L;leukocytes:adults–4000–10,000cell/mL;hemoglobin:adults–men14–18g/dL,women12–16g/dL; hematocrit:adults–men41–52%,women36–46%;platelets:adults–150,000–400,000cells/mL.
CD4Tcellcount(p=0.15),HIVviralload(0.37),LDH(0.09),BG (0.08),andtimesinceHIVdiagnosis(p=0.45).
GMwasanalyzedinserumof56patientsandinBALoftwo patients.OnepatienthadserumandBALpositives(PJP)and onepatienthadpositiveserumsample(CAP).AlltheotherGM testswerenegative.
Results of univariate analysis of the three groups of patientsaresummarizedinTable2.ThemeanvaluesofBG in the group of patients clinically diagnosed as P. jirovecci (19 patients; 240.8±185.7pg) was significantly higher than in patients with CAP(18 patients; 36.3±34.2pg;p<0.0001) and in those diagnosed withother infections (23 patients; 67.3±60.7pg;p<0.0001). Eightpatients notdiagnosed with PJPhadBG>80pg/mL(histoplasmosis:1;strongyloidiasis:1; tuberculosis: 1; cryptococcosis: 1; CAP: 2; undiagnosed: 2). ConsideringtheBGcut-offof>80pg/mL,thesensitivityand specificity of BG in the PJP patients were 0.90 and 0.80, respectively. Two patients withPJP had BG<80pg/mL.LDH wasalso significantlyhigher inthe group ofpatients with P. jirovecii (mean 762.47±433.18U/L) than in patients with CAP(379.5±99.9U/L),andotherdiagnosis (442.6±217.6U/L) (p=0.003). Patients with PJP had lower T CD4+lymphocyte count (p=0.001) and duration of time since HIV diagnosis (p=0.001).However,higherBGandLDHweretheonly indepen-dentvariablesassociatedwiththediagnosisofPJPinrelation tothediagnosisofCAPandotherinfections(Table3).
Eleven(68.7%)ofthe16patientswhowerefirstdiagnosed asHIVduringhospitalization hadPJP (ninepatients)orPJP associatedwithMycobacteriumnon-tuberculosis(twopatients) andofthe otherfive,twopatientshadhistoplasmosis,one cryptococcosis,onetuberculosis,andonepatienthadLRTI.
Tuberculosis(17cases),Pneumocystosis(16cases),Herpes zoster(13cases),cerebraltoxoplasmosis(10cases)and hepati-tisC(11cases)werethemostprevalentpreviousopportunistic infectionsreportedinourpatients.Sixpatientsincludedin thisstudyhadapreviousdiagnosisofPJP.Threepatientsoutof the17previouslydiagnosedwithtuberculosishadatreatment failureorrelapsed.
Thirty-fivepatientshadatleastoneclinicalspecimen(BAL: 1,sputum:14;OL:13;sputum+OL:6;sputum+BAL:1)tested byLAMP.ThismolecularmethoddetectedP.jiroveciiDNA prod-uctinclinicalspecimenoffivepatients(1BAL;4sputum).All
patientswhoseLAMPturnedoutpositivehadchestX-rayor CTsuggestiveofPJPandBG≥341.55pg.LAMPwerepreformed in12patientswithPJPandonlyinfiveitturnedoutpositive.
Discussion
Pulmonary infectionisoneofthemostfrequent complica-tion inpeople livingwithHIV/AIDS, remainingthe leading cause ofmorbidityandmortalityworldwide.7 Inourstudy,
pneumocystosis, CAP,tuberculosis,andLRTIwerethemost frequentcausesofhospitalization.Ofnote,fungalinfections represented36.6%(22patients)ofourcases.Pneumocystosis wasthemostfrequentAIDS-definingopportunisticinfection representing 68.7% ofnewly diagnosed AIDS cases, under-scoringtheneedtoimproveearlydetectionofHIVinfection. Remarkably, tuberculosiswas AIDS-defining illness inonly onepatient.
Systematicreviewshavedemonstratedthatopportunistic infections decreasedaftertheavailabilityofART,mostlyin high-income countries.21 However,there isa lackof
infor-mation on the epidemiology of opportunistic infections in low-andmiddle-incomecountries,includingBrazil.Ourstudy demonstratedthatopportunisticinfectionswerestill preva-lent. Astudy from San Francisco reported that PJP (39.1%) followedbyKaposisarcoma(20.1%)werethemostcommon AIDS-definingdiseases.22Grinsztejnetal.23found
tuberculo-sisasthemostcommonAIDS-definingopportunisticinfection in a Brazilian cohort compared to esophageal candidiasis amongUSHIV-infectedpatients.AnotherstudyfromRiode Janeiro,concludedthattuberculosiswasthe mostfrequent AIDS-defining opportunistic illness,24 whereas in a recent
series of HIV-infected patients from a referral hospital in Brazil,pneumocystosiswasstilltheprevailingopportunistic infectionintheHIVgroup.25Thedifferencebetweentherates
ofopportunisticinfectionsfoundinourresultscomparedto otherstudiescouldbeexplainedbythecausesofpulmonary infectionswehavestudiedandnotallcausesof opportunis-ticinfections inAIDSpatients. Notably, mostinternational andnationalstudiesarefromsinglecenters.Althoughtheir resultscontributedtotheunderstandingofthelocal epidemi-ology,theconductofmulticenterstudieswillleadtobetter
Table2–Univariateanalysisofthe(1-3)--D-Glugan,LDH,HIVviralload,CD4Tcellcount,durationofHIVinfection,and outcomeofthethreegroupsofpatientsincludedinthestudywithpulmonaryinfections.
Variables Clinicaldiagnosis p-Value
PJP(n=19) CAP(n=18) Other(n=23) (1-3)-ˇ-D-Gluganinserum(pg/mL) Median 183 29.8 52.8 Mean(±SD) 240.8(±185.7) 36.3(±34.2) 67.3(±60.7) <0.0001 LDH(U/L) Median 761 419 441 Mean(±SD) 762.47(±433.18) 379.5(±99.9) 442.6(±217.6) 0.003
Viralload(copies/mL)
Median 62,609 3800 74,892
Mean(±SD) 801,171(±2,194,157) 149,525(±270,015) 482,859(±135,562) 0.30
LymphocyteTCD4+(cell/mm3)
Median 40 230 97
Mean(±SD) 73(±107) 303(±267) 217(±312) 0.001
TimesincediagnosisofHIVinfection(years)
Median 0 15 11
Mean(±SD) 4.94(±6.83) 14(±7.67) 10.7(±8.32) 0.001
Outcome
Discharge 15 15 21 0.62
Death 4 3 2
Meantimeofhospitalization(indays) 22.9(±22.8) 13.4(±8.9) 16.1(±12.3) 0.31 PJP,Pneumocystisjiroveciipneumonia.PJPgroupincluded:P.jirovecii+communityacquiredpneumonia:4patients;P.jirovecii+Mycobacterium non-tuberculosis:2patients.CAP:communityacquiredpneumonia.Other:lowerrespiratoryinfection:9patients;tuberculosis:5patients; histo-plasmosis:2patients;cryptococcosis:1patient;disseminatedstrongyloidiasis:1patient;nocardiosis:1patient;pulmonaryembolism:1patient andundiagnosed:3patients.
Table3–MultivariatelogisticregressionanalysisaofvariablesindependentlyassociatedwithP.jiroveciipneumonia,
community-acquiredpneumonia,andotherpulmonarydiseases.
Variable Group p-Value OR CI95%
(1-3)--D-Glugan(pg/mL) PJPvs.Other 0.0267 1.015 1.002;1.028
(1-3)--D-Glugan(pg/mL) CAPvs.Other 0.1028 0.986 0.970;1.003
(1-3)--D-Glugan(pg/mL) PJPvs.CAP 0.0051 1.029 1.009;1.050
LDH(U/L) PJPvs.Other 0.0375 1.009 1.000;1.017
LDH(U/L) CAPvs.Other 0.3617 0.997 0.992;1.003
LDH(U/L) PJPvs.CAP 0.0154 1.011 1.002;1.021
PJP,P.jiroveciipneumonia;CAP,community-acquiredpneumonia;Other,otherpulmonarydiseases;LDH,lactatedehydrogenase;OR,OddsRatio; CI,confidenceinterval.
a Astheanalyzedvariableswerecontinuous,theresultsofthemultivariatelogisticregressionanalysisdescribedtheoddsratiovaluesasthe effectoftheincrementofeach“1”unitofBG(pg/mL)orLDL(U/L)levels,intheincreaseofthechanceofapatientbeinginthePJPgroup.
preventionand treatmentstrategies.One limitationofthis studywasthefactthatonly60AIDSpatientsmetthe inclu-sioncriteriaduringthestudyperiod.Thetimeframeofthe study,thelimitednumberofbedsintheinfectiousdiseases ward,andthelonglengthhospitalstaywereexternalfactors thatcontributedinreducingthesamplesizeofourstudy.
Clinicianssearchfornon-invasivemethods,suchasserum markers,asanaidtodiagnosepulmonaryinfiltrates, partic-ularly in HIV/AIDS. BG values were considerably higher in patientsclinicallydiagnosedwithPJP and mightbeagood marker forpatients with AIDS and pulmonary symptoms. ThevisualizationoftheP. jiroveciiinpulmonaryspecimens isdifficult,whereasthedosageofserumBGcouldbea reli-able marker in the diagnosis of PJP.10 Considering the BG
cut-offvaluerecommendedbythemanufacturer(80pg/mL), we found a high sensitivity(0.90) and specificity (0.80). In two other studies, that included a higher number of HIV patientsthanourstudy,theauthorsfound,usingthesame cut-off,asensitivityandspecificityof0.91/0.64and0.98/0.39, respectively.10,26OnepotentiallimitationoftheuseofBGin
clinicalpracticewithHIV-infectedpatientsmightbethe posi-tivityofthetestinotherinvasivemycosis,suchascandidiasis, fusariosis,aspergillosis,andhistoplasmosis.27,28
PreviousstudieshavefoundthathighlevelsofLDHwere associatedwithPJP,andthatLDHcouldbeaserummarker inthediagnosisofpneumocystosis.10,29,30Ourdatashowed
a significant correlation between high levels of LDH and pneumocystosis(p=0.003)(Table2).Themultivariateanalysis
showed that higher levels of BG and LDH were indepen-dentlyassociatedwithPJPwhencomparedwiththeresultsin patientswithCAPandotherpulmonaryconditions(Table3). Inthemultivariate logisticregression analysisdescribed in Table3theoddsratiosrepresenttheeffectoftheincrementof each“1”unitofBG(pg/mL)orLDL(U/L)levels,inthechance ofapatienthavingPJP, asthesevariables werecontinuous variables.31Basedinourresults,thecombinationofmorethan
oneserummarkermaybehelpfulinthediagnosisofPJPand todiscriminatePJPfromotherpulmonarydiseases.
LAMPisarapidsimpleandcost-effectivemethodfor diag-nosingvariouspathogensincludingPneumocystis.9LAMPforP.
jiroveciiwastestedintherespiratorysecretionsofourpatients andtheresultswerenotencouraging.Mostoftheclinical sam-plesfromourpatientsweresputumorOL,andperhapsifwe hadtestedBAL,wewouldhaveobtainedbetterresults. How-ever,inclinicalpractice,non-invasivediagnosticmethodsare preferredoverinvasiveprocedures,suchasBALorbiopsies.In ourseriesofcases,serumBGandLDHprovidedbetteraidfor thediagnosisofpneumocystosisthanLAMP.
Inconclusion,ourresultsshowedthatPJPstillhasahigher prevalenceoverotherpulmonaryconditionsinourgeographic area,and thatweneedtoimplementpromptandaccurate diagnostictestsforpulmonaryinfections.BGandLDHshowed to be suitable serum markers in the diagnosis of PJP and shouldbeimplementedinroutinecaretoassistourHIV/AIDS population.Moreover,mostofourpatientswere previously enrolled in a HIV/AIDS care center. Although, we have a well-establishedNationalprogramforHIV/AIDSinBrazil,the HIVcascade care isneglecting the early diagnosis,patient follow-up,inadditiontoinsufficientpatientadherencetothe HIV/AIDSprogram. Thisstudy showedthat thereisaneed forbetterstrategiestoachieveearlydiagnosisofHIVinfection andthereforereducingopportunisticinfectionsintheinfected patients.
Conflicts
of
interest
Theauthorshavenoconflictsofinterest.
Acknowledgements
This project was supported by grants from the collabo-rative research project: Science and Technology Research Partnership for Sustainable Development, Japan (SATREPS) and University of Campinas, Brazil, No. 02P-29548-09 and Fundac¸ão de Amparo à Pesquisa do Estado de Sao Paulo, FAPESPNo.2012/51158-0.WearegratefultoMr.RogerTimothy RentfrowfortheEnglishproofreading.
r
e
f
e
r
e
n
c
e
s
1. WHO.Globaltuberculosisreport2015,20thedition. http://apps.who.int/iris/bitstream/10665/191102/ 1/9789241565059eng.pdf?ua=1.
2. HullMW,PhillipsP,MontanerJS.Changingglobal epidemiologyofpulmonarymanifestationofHIV/AIDS. Chest.2008;134:1287–98.
3.CrothersK,ButtAA,GilbertCL,etal.IncreasedCOPDamong HIV-positivecomparedtoHIV-negativeveterans.Chest. 2006;130:1326–33.
4.BowerM,PowlesT,NelsonM,etal.HIV-relatedlungcancerin theeraofhighlyactiveantiretroviraltherapy.AIDS.
2003;17:371–5.
5.MurrayJF.Epidemiologyofhumanimmunodeficiency virus-associatedpulmonarydiseases.ClinChestMed. 2013;34:165–79.
6.BenitoN,MorenoA,MiroJM,TorresA.Pulmonaryinfections intheHIV-infectedpatients:anupdateinthe21stcentury. EurRespirJ.2012;39:730–45.
7.LazarousDG,O’DonnellAE.Pulmonaryinfectioninthe HIV-infectedpatientintheeraofhighlyactiveantiretroviral therapy:anupdate.CurrInfectDisRep.2007;9:228–32. 8.Armstrong-JamesD,MeintjesG,BrownGD.Aneglected
epidemic:fungalinfectionsinHIV/AIDS.TrendsMicrobiol. 2014;22:120–7.
9.SinghP,SinghS,MirdhaBR,etal.Evaluationof loop-mediatedisothermalamplificationassayforthe detectionofPneumocystisjiroveciiinimmunocompromised patients.MolBiolInt.2015;2015:819091.
10.EstevesF,LeeCH,deSousaB,etal.(1-3)-Beta-D-glucanin associationwithlactatedehydrogenaseasbiomarkersof Pneumocystispneumonia(PCP)inHIV-infectedpatients.EurJ ClinMicrobiolInfectDis.2014;33:1173–80.
11.SalermoD,MushattD,MyersL,etal.SerumandBAL beta-D-glucanforthediagnosisofPneumocystispneumoniain HIVpositivepatients.RespirMed.2014;108:1688–95.
12.TasakaS,KobayashiS,KazumaY,etal.Serum(1-3) -D-glucanassayfordiscriminationbetweenPneumocystis jiroveciipneumoniaandcolonization.JInfectChemother. 2014;20:678–81.
13.KarageorgopoulosDE,QuJM,KorbilaIP,etal.Accuracyof beta-D-glucanforthediagnosisofPneumocystisjirovecii pneumonia:ameta-analysis.ClinMicrobiolInfect. 2013;19:39–49.
14.FinkelmanMA.Pneumocystisjiroveciiinfection:Cellwall (1-3)-B-D-glucanbiologyanddiagnosticutility.CritRev Microbiol.2010;36:271–81.
15.SaxPE,KomarowL,FinkelmanMA,etal.AIDSClinicalTrials GroupStudyA5164Team:blood(1→3)-{beta}-D-glucanasa diagnostictestforHIV-relatedPneumocystisjirovecii pneumonia.ClinInfectDis.2011;53:197–202.
16.LimWS,EerdenMM,LaingR,etal.Definingcommunity acquiredpneumoniaseverityonpresentationtohospital:an internationalderivationandvalidationstudy.Thorax. 2003;58:377–82.
17.WHOcasedefinitionsofHIVforsurveillanceandrevised clinicalstagingandimmunologicalclassificationof HIV-relateddiseaseinadultsandchildren.HIV/AIDS Programme2006.
18.CorrêaRA,LundgrenFLC,Pereira-SilvaJL,etal.Diretrizes brasileirasparapneumoniasadquiridasnacomunidadeem adultosimunocompetentes[Brazilianguidelinesfor community-acquiredpneumoniainimmunocompetent adults–2009].JBrasPneumol.2009;35:574–601.
19.UemuraN,MakimurK,OnazakiM,etal.Developmentofa loop-mediateisothermalamplificationmethodfordiagnosing Pneumocystispneumonia.JMedMicrobiol.2008;57:50–7. 20.DeanAG,SullivanKM,SoeMM.OpenEpi:OpenSource
EpidemiologicStatisticsforPublicHealth,Version. www.OpenEpi.com[updated04.05.15/accessed18.04.16]. 21.CoelhoL,VelosoVC,GrinsztejnB,etal.Trendsinoverall opportunisticillnesses,Pneumocystiscariniipneumonia, cerebraltoxoplasmosisandMycobacteriumaviumcomplex incidenceratesoverthe30yearsoftheHIVepidemic:a systematicreview.BrazJInfectDis.2014;18:196–210.
22.DjaweK,BuchaczK,HsuL,etal.Mortalityriskafter AIDS-definingopportunisticillnessamongHIV-infected persons-SanFrancisco,1981–2012.JInfectDis.
2015;212:1366–75.
23.GrinsztejnB,VelosoVG,PilottoJH,etal.Comparisonof clinicalresponsetoinitialactiveantiretroviraltherapyinthe patientsinclinicalcareIntheUnitedStatesandBrazil. JAcquirImmuneDeficSyndr.2007;45:515–20.
24.CoelhoL,CardosoSW,AmancioRT,etal.Trendsin
AIDS-definingopportunisticillnessesincidenceover25years inRiodeJaneiro,Brazil.PLOSONE.2014;9:e98666.
25.GalisteuKJ,CardosoLV,FuriniAAC,etal.Opportunistic infectionsamongindividualswithHIV-1/AIDSinthehighly activeantiretroviraltherapyeraataquaternarylevelcare teachinghospital.RevInstMedTrop.2015;48:149–56. 26.SalermoD,MushattD,MyersL,etal.SerumandBAL
beta-D-GlucanforthediagnosisofPneumocystispneumoniain HIV-positivepatients.RespirMed.2014;108:1688–95.
27.AzoulayE,guigueN,DarmonM,etal.(1,3)--D-glucanassay fordiagnosinginvasivefungalinfectionsincriticallyill patientswithhematologicalmalignancies.Oncotarget.2016 [Feb18,Epubaheadofprint].
28.HeS,HangJP,ZhangL,etal.Asystematicreviewand meta-analysisofdiagnosticaccuracyofserum1,3--D-glucan forinvasivefungalinfection:focusoncutofflevels.
JMicrobiolImmunolInfect.2015;48:351–61.
29.VogelM,WeissgerberP,GoeppertB,etal.Accuracyofserum LDHelevationforthediagnosisofPneumocystisjirovecii pneumonia.SwissMeWkly.2011;141:w13184. 30.QuistJ,HillAR.Serumlactatedehydrogenase(LDH)in
Pneumocystiscariniipneumonia,tuberculosis,andbacterial pneumonia.Chest.1995;1085:415–8.
31.HosmerDW,LemeshowS.Interpretationofthecoefficientsof thelogisticregressionmodel.In:Appliedlogisticregression. NewYork:JohnWiley&SonsInc.;1989.p.38–81.