Review
Pulmonary nocardiosis in Western Europe — Clinical evaluation of 43 patients and population-based estimates of hospitalization rates
Sebastian R. Ott
a,b, N. Meier
a,b, Martin Kolditz
c, Torsten T. Bauer
d,*, Gernot Rohde
e, Elisabeth Presterl
f, Dirk Schürmann
g, Philipp M. Lepper
h, Felix C. Ringshausen
i, Holger Flick
j, Stephen L. Leib
k, Mathias W. Pletz
l, for the OPINION Study Group
aDepartmentofPulmonaryMedicine,UniversityHospital(Inselspital)andUniversityofBern,Bern,Switzerland
bDepartmentofPulmonaryMedicineandThoracicSurgery,St.Claraspital,Basel,Switzerland
cDivisionofPulmonology,MedicalDepartment1,UniversityHospitalofTUDresden,Dresden,Germany
dDepartmentofPneumology,LungenklinikHeckeshorn,HeliosKlinikumEmilvonBehring,Berlin,Germany
eDepartmentofPneumologyandAllergology,GoetheUniversity,Frankfurt,Germany
fKlinischesInstitutfürKrankenhaushygiene,UniversityofVienna,Vienna,Austria
gDepartmentofInternalMedicine/InfectiousDiseasesandPulmonaryMedicine,Charité-UniversitätsmedizinBerlin,Berlin,Germany
hDepartmentofInternalMedicineV,UniversityHospitalofSaarland,Homburg,Germany
iDepartmentofRespiratoryMedicine,HannoverMedicalSchool,Hannover,andGermanCentreforLungResearch(DZL),Hannover,Germany
jDivisionofPulmonology,DepartmentofInternalMedicine,MedicalUniversityofGraz,Graz,Austria
kInstituteforInfectiousDiseases,UniversityofBern,Bern,Switzerland
lCentreforInfectiousDiseasesandInfectionControl,JenaUniversityHospital,Jena,Germany
ARTICLE INFO Articlehistory:
Received18July2018
Receivedinrevisedform12December2018 Accepted21December2018
CorrespondingEditor:EskildPetersen,Aar- hus,Denmark
Keywords:
Nocardiosis Nocardia
Pulmonarynocardiosis
ABSTRACT
Background:Pulmonarynocardiosis(PN)isanuncommonbutpotentiallylife-threateninginfection.Most of our knowledgeonPN isderivedfrom casereportsand small caseseries. Increasingincidencerates ofPN have beenreportedrecently.TheaimofthisstudywastodescribetheclinicalcourseofandriskfactorsforPNin fourWesternEuropeancountriesandtoestimatepopulation-basedannualhospitalizationrates.
Methods:Thiswasaretrospectiveevaluation(1995–2011)oftheclinicalcourseofandriskfactorsforPNin patientsat11hospitalsinfourEuropeancountries(Germany,Austria,Switzerland,andtheNetherlands).
Population-basedestimatesofhospitalizationratesforPNinGermany(2005to2011)werecalculated usingofficialGermannationwidediagnosis-relatedgroups(DRG)hospitalstatistics.
Results:Forty-threepatientsfulfilledstringent criteriaforproven(n=8)andprobable(n=35)PN;sevenhad extrapulmonarydissemination.Forthese43patients,themajorriskfactorsforPNwereimmunocom- promising(83.7%)and/orpulmonary(58.1%;asonlycomorbidityin27.9%)comorbidities.Themedian durationofPNtargetedtherapywas12weeks.Distinctivepatternsofresistancewereobserved(imipenem susceptibility:Nocardiafarcinica33.3%;Nocardiaasteroides66.7%).Theoverallmortalityratewas18.9%
(50%indisseminatedPN).Overtime,annualPNhospitalizationratesremainedunchangedataround0.04/
100000,withthehighestrateamongmenaged75–84years(0.24/100000).
Conclusions:PNisarare,butpotentiallylife-threateningdisease,andmainly affectsimmunocompromised elderlymales.Overall,annualhospitalizationratesremainedstablebetween2005and2011.
©2019TheAuthors.PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc- nd/4.0/).
Contents
Introduction ... 141 Materialsandmethods ... 141 Cohortstudy ... 141
*Correspondingauthorat:RespiratoryDiseasesClinicHeckeshorn,HeliosClinicEmilvonBehring,Walterhoeferstr.11,D-14165Berlin,Germany.
E-mailaddress:torsten.bauer@helios-gesundheit.de(T.T. Bauer).
https://doi.org/10.1016/j.ijid.2018.12.010
1201-9712/©2019TheAuthors.PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).
ContentslistsavailableatScienceDirect
International Journal of Infectious Diseases
j o u r n a lh o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / i j i d
Casedefinition ... 141
Microbiology ... 141
Definitions ... 142
EstimationofthehospitalizationrateinGermany ... 142
Statisticalanalysis ... 142
Results ... 142
Cohortstudy ... 142
Patients ... 142
Clinicalandradiologicalfindings ... 142
Microbiologicalfindings... 143
Treatmentandoutcome ... 144
Population-basedestimatesofannualratesofhospitalizedPNinGermanybetween2005and2011 ... 144
Discussion ... 144
Acknowledgement ... 147
References... 147
Introduction
Nocardiosis is a rare infection caused by Nocardia spp.
Nocardiaceae are aerobic, Gram-positive, branching, partially acid-fast filamentary bacteriathat arefoundworldwide insoil, decomposingvegetation,dust,andinfreshaswellassaltwater (Brown-Elliottetal.,2006).Morethan60differentspecieshave been identified and half of them have been described as pathogens in humans (Martinez et al., 2008). Despite their ubiquitousoccurrence,Nocardiaspp.areconsideredtobeneither a commensal of the normal human flora nor a common laboratory contaminant (McNeil and Brown,1994). Isolationof Nocardiaspp.fromlowerrespiratorytractsamplesinthepresence of clinical signs and symptoms of a lower respiratory tract infectionorcompatibleradiographicfindingsshouldalwaysraise suspicion of a causative role for this pathogen. Pulmonary nocardiosis(PN)mayhavea potentiallylife-threateningcourse, especially in disseminated cases (Hardak et al., 2012; Beaman et al., 1976; Boiron et al., 1992; Minero et al., 2009). So far, transmissionfromhumantohumanhasnot beendocumented, and dissemination usually results from a pulmonary focus (Ambrosionietal.,2010).
Nocardiosis mainly manifests as an opportunistic infection in immunocompromisedpatients(Chenetal.,2013).Animpairmentof cell-mediated immunity isconsidered to be the major risk factor (Tremblayetal.,2011;Deemetal.,1983;BeamanandBeaman,1994;
FiliceandNiewoehner, 1987).Takingintoaccounttheincreasinguseof immunosuppressive therapies, including new treatments such as biological medicines, e.g. monoclonal antibodies, the incidence of Nocardiainfectionseemstobeontherise(Mineroet al.,2009;
Tremblay et al., 2011).However, reports of immunocompetent patientswithnocardiosisarealsoemerging(Wilson,2012).Inthese patients, pulmonary comorbidities, mainly chronic obstructive pulmonary disease (COPD), are important risk factors for PN (Martinezetal.,2008).
Theaimsofthisstudywere(1)tocharacterizetheriskfactors, microbiology,andclinicalcourseofPNinawell-definedcohortof patientsinfourWesternEuropeancountries,and(2)toestimate population-basedage-and sex-specific,aswellas age-adjusted hospitalizationratesforPNinGermany.
Materialsandmethods
Cohortstudy
Thisretrospectivemulticentrestudyincludedpatientsfrom11 hospitals in four countries (Germany (n=7), Austria (n=2), Switzerland (n=1), the Netherlands (n=1)); the patients were diagnosed withPN betweenJanuary1,1995 and December 31,
2011. A search of the microbiological laboratory data at each institutionwasperformedtoidentifyeligiblepatients.Datawere collectedfromthepatientmedicalrecords,includingdemographic characteristics,clinicalsigns and symptoms,radiologicalabnor- malities on chest X-ray or computed tomography (CT) scans, laboratoryresults,microbiologicalfindings,comorbiditiesandrisk factors, underlying pulmonary morbidities, immunosuppressive medications,antibiotictreatment,andoutcome.
Casedefinition
OnlypatientswithaprovenorprobablepulmonaryNocardia infection were included. Due to clinical similarities to fungal infections,provenandprobablePNweredefinedaccordingtothe European Organisation for Research and Treatment of Cancer (EORTC)criteriaforinvasiveaspergillosis.Patientshadtofulfilone ofthefollowingconditionsforstudyentry:
(1) ProvenPN:apositiveculturefromsamplestakenfromasterile primarysite,e.g.pleuraleffusion,blood,Cerebrospinalfluid, tissuebiopsy,orabscesspunctureAND(i)radiologicalfindings on chest X-ray or CT scan of the thorax compatible with pulmonarynocardiosisOR(ii)clinicalsignsandsymptomsofa lowerrespiratory tract infection (cough, sputum,dyspnoea, fever>38.3C).
(2) ProbablePN(oneofthefollowing):(i)apositiveculturefrom sputum,tracheobronchialaspirate,orbronchoalveolarlavage (BAL) AND (a) clinical signs and symptoms of a lower respiratory tract infection (cough, sputum, dyspnoea, fever
>38.3C)AND(b) radiologicalfindings onchestX-rayorCT scanofthethoraxconsistentwithpulmonarynocardiosis;(ii) microscopicproofofGram-positive,branching,partiallyacid- fastfilamentarybacteriainhistologicalsamples(identifiedas Nocardiaspp.)AND(a)clinicalsignsandsymptomsofalower respiratory tract infection (cough, sputum, dyspnoea, fever
>38.3C)OR(b)radiologicalfindingsonchestX-rayorCTscan ofthethoraxconsistentwithpulmonarynocardiosis.
PatientswithprovenandprobablePNwerethenincludedasa singlecohortandtheirdatasummarizedandanalyzed.Patients undertheageof18yearswereexcludedfromthisstudy.
Microbiology
At all institutions, the identification of Nocardia spp. and susceptibilitytestingwereperformedonlyuponspecialrequest.
Since genotypic methods for species identification as part of clinicalroutine(e.g., 16SrRNAgenePCR)becameavailableonlylate duringthestudyperiod, patientswithconventionalphenotypic andbiochemical speciesidentificationwereincluded,aswellas patientswithgenotypicidentificationresults.
Definitions
AdisseminatedcoursewasdefinedasprovenorprobablePN withinvolvementofatleastoneothernon-contiguousorgan,ora radiologicalfinding,e.g.abscessformationinorgansotherthanthe lungs, such as the brain or abdomen, interpreted as being compatiblewithNocardiainfectioninapatientwithPN.
Immunosuppressive medicationwas definedas chronic (>4 weeks)intake ofsteroidsat a dosageof>10mgprednisone or equivalentper day, or the use of another immunosuppressive agentsuchasacalcineurininhibitor,mycophenolate,cytostatic agent,mTORinhibitor, TNF-bindingprotein,or otherbiological medicine.
EstimationofthehospitalizationrateinGermany
Fortheestimationoftheannualpopulation-basedhospitaliza- tionratesforPN,theofficialGermannationwidediagnosis-related groups(DRG)hospitalstatisticswereanalyzed(GermanFederal HealthMonitoringSystem, 2013).These have includeddata on primarydischargediagnosessinceinceptionandonallsecondary dischargediagnosessince2005,fromallGermanhospitalsusing DRGbillingofmedicalservices.AlmostallGermanhospitalsuse this system, with the exception of hospitals for prevention, rehabilitation, and mental and mood disorders, and day care units.Caseswereidentifiedbyextractingallrecordsforwhichthe four-digit International Classification of Diseases,10th revision (ICD-10)code A43.0(pulmonary nocardiosis) withinthe three- digit category A43 (nocardiosis) was listed as a primary or secondaryhospitaldischargediagnosis.De-identifiedDRGdiag- nosisdatawereprovidedforthewholeofGermanyasabsolute numbersstratifiedbyagegroup,sex,andyearofdiagnosis.
Statisticalanalysis
Resultsareexpressedasfrequenciesorasthemeanstandard deviation(SD)unlessindicatedotherwise.Thetwo-tailedFisher’s exacttestwasusedtocompareproportions.Thesignificancelevel wassetto5%forallanalyses,andp-valuesarereported.Alldata wereanalyzedand processed usingIBM SPSSStatistics version 21.0(IBMCorp.,Armonk, NY,USA)ona WindowsXPoperating system(Microsoft,Redmond,WA,USA).
For the estimation of PN hospitalization rates, all hospital- izationswithPNastheprimaryorsecondaryhospitaldischarge diagnosisin Germany from 2005 through 2011 were included.
OfficialGermancensusage-andsex-specificpopulationdatawere used as the denominator for all calculations (German Federal HealthMonitoringSystem,2013).Ageadjustmentwasperformed bydirectmethodinordertocontrolfordifferentagedistributions acrossGermany andtoallow forcomparison betweendifferent years.Age-adjustedhospitalizationrateswereestimatedusingthe appropriateGermanCensusStandardPopulationasthereference population.ThedataanalysiswasperformedinExcel(Microsoft).
Results
Cohortstudy Patients
Fifty-eightpatientswithmicrobiologicallyprovenNocardiaspp.
were identified. Of these, 43 patients fulfilled the criteria for proven(n=8)or probable(n=35)PN andwereincludedinthe study.MostpatientswererecruitedinGermany(n=22),followed byAustria(n=13),Switzerland(n=6),andtheNetherlands(n=2).
Fortypatients(93.0%)presentedwithrelevantcomorbidities.A pulmonarycomorbiditywasthemostfrequentpredisposingfactor (n=25); this was the only underlying comorbidity in 12/43
patients (27.9%). Twenty-two patients (51.2%; n=22/43) were receiving immunosuppressive treatment, of whom eight were receiving corticosteroid monotherapy(see Table 1). HIVtesting wasperformedfor20patientsandalltestednegative.Twenty-four patients had received antibiotic treatment prior to admission.
Detailsofthedemographicandepidemiologicalcharacteristicsare summarizedinTable1.
Clinicalandradiologicalfindings
Theclinicalpicturewasnon-specificandcomparabletoother entitiesoflowerrespiratorytractinfection(seeTable2).Dueto incompletedata, information onsputum colour,pleuritic chest pain,andhaemoptysiswerenotavailable.Overall,inflammatory parameterswereonlymoderatelyelevated,despitewiderangesof values(Table2).Owingtothenon-specificclinicalappearanceof PN,themediantimetofinaldiagnosiswas 14days(range0–54 days).
Detailedclinical and/orradiological dataonextrapulmonary involvementwere available for 38 patients, and sevenpatients (18.4%;n=7/38)sufferedfromdisseminatednocardiosis(Table2).
Thecentralnervoussystem(CNS)wasthemostfrequentlyaffected site(n=5).Twopatientspresentedextrapulmonaryfociinmore thanoneadditionalorgan.
Table1
Demographicandepidemiologicalcharacteristicsof43patientswithpulmonary nocardiosis.a
Age(years),meanSD(range) 55.814.4
(21–75) Sex
Male 31/43(72.1%)
Female 12/43(27.8%)
Admissiondiagnosis(n=38)
Pneumonia 21/38(55.3%)
Lungcancer/metastasis 3/38(7.9%)
Tuberculosis 2/38(5.3%)
Acuteexacerbationofapre-existinglungdisease(e.g.
bronchiectasis,COPD)
2/38(5.3%)
Pleuraleffusion/empyema 2/38(5.3%)
Pulmonarynocardiosis 2/38(5.3%)
Otherb 6/38(15.8%)
Antibioticpretreatment 24/43(55.8%)
Underlyingdiseasec 40/43(93.0%)
Pulmonarymorbidityc 25/43(58.1%)
COPD 8/43(18.6%)
Tuberculosis 5/43(11.6%)
Bronchiectasis 3/43(7.0%)
Interstitiallungdisease 4/43(9.3%)
Otherpulmonarydiseased 10/43(23.3%)
Solidorgantransplantation 11/43(25.6%)
Haematologicaldisease 10/43(23.3%)
Diabetesmellitus 6/43(14.0%)
Autoimmunedisease 5/43(11.6%)
Liverdisease 2/43(4.7%)
Severerenalimpairment 2/43(4.7%)
Immunosuppressivetherapy 22/43(51.2%)
Combinationtherapy 14/43(32.6%)
Monotherapy(corticosteroid) 8/43(18.6%)
COPD,chronicobstructivepulmonarydisease;CMV,cytomegalovirus;SD,standard deviation.
a Resultsarepresentedasthenumber andpercentage(%),unlessspecified otherwise.
b Onepatienteachwithoneofthefollowing:haemoptysis,feverofunknown origin,meningitis,vertigoandnausea,rejectionfollowinglungtransplantation, cryptogenicorganizingpneumonia.
cPatientscouldhavemorethanoneunderlyingdiseaseorpulmonarymorbidity.
d One patient each with: asthma, cystic fibrosis, chronic pulmonary CMV infection, obstructive sleep apnoea, pulmonary aspergillosis, non-tuberculous mycobacteriainfection,primaryciliarydyskinesia,pulmonaryp-ANCAvasculitis, chronicbronchitis.
EvaluablechestX-rays and/orCTscanswereavailablefor39 patients.Similartotheclinicalpicture,radiologicalfindingswere alsovariable,non-specific,andheterogeneous.Themostfrequent abnormalitywaspulmonaryairspaceconsolidation(89.7%;n=35/
39) withor without air bronchograms, followed by (reticulo)- nodular changes (30.1%; n=12/39). Abscess-forming consolida- tionsorcavitationswerepresentinninepatients(23.1%;n=9/39) andpleuraleffusioninfourpatients(10%;n=4/39).Detailsofthe distributionofradiologicalfindingsarepresentedinFigure1.
Microbiologicalfindings
Dataonspeciesidentificationwereavailablefor28patients.A totalof37differentisolateswereidentified(seeFigure2).Asingle Nocardiaspp.(n=6)wasidentifiedinmostpatients;fourpatients hadtwodifferentspecies,onepatienthadthreedifferentspecies, and one patient had four different species. Nocardia farcinica (n=8), Nocardia asteroides (n=7), and Nocardia cyriacigeorgica (n=3)accountedforalmost50%ofallinfectionswithidentification beyondjustgenusrecognition.Dataonthedurationofgrowthin culturewere available for 24patients, and the mediantime to growthwas16days(mean18.214.7days;range2–58days).
Specimens obtained by bronchoscopy such as BAL, trans- bronchial lung biopsy, and tracheobronchial aspirate were diagnosticin27patients, indicatingthatendoscopicprocedures were performed in approximately two-thirds of all patients (Table 3). Nineteen additional microorganisms were recovered from16patients(37.2%;n=16/43)(Table3).
Table2
Clinicalsignsandsymptoms,additionalextrapulmonaryfoci,timefromadmission todiagnosis,andlaboratoryvalues.
Number %
Clinicalsignsandsymptomsa(n=43)
Cough 33/43 76.7%
Fever 27/43 62.8%
Sputum 26/43 60.5%
Dyspnoea 17/43 39.5%
Weightloss 7/43 16.3%
Disseminatedcourse,extrapulmonaryfocia (n=38)
7/38 18.5%
Brain 5/38 13.2%
Softtissue 2/38 5.3%
Skin 1/38 2.6%
Abdomen 1/38 2.6%
MeanSD Median (range) Timefromadmissiontofinaldiagnosis(days)
(n=29)
16.813.2 14.0(0–54) Inflammatorymarkersonbloodexamination
(n=37)
CRP(mg/l) 102.1104.9 66.0(0–429)
WBC(109/l) 11.68.5 10.5(2.1–
37.1) CRP,C-reactiveprotein;SD,standarddeviation;WBC,whitebloodcellcount.
aPatientscouldhavemorethanone.
Figure1.Radiologicaldistributionofabnormalitiesin39patientswithpulmonarynocardiosis.
Evaluabledataonantimicrobialsensitivitywerelimitedto20 isolatesfromsixdifferentspecies(Table4).
Treatmentandoutcome
All patients received antibiotic therapy, and a variety of antibacterial agents were given (see Table 5). Orally and intravenously applied trimethoprim–sulfamethoxazole (TMP– SMX; 15mg/kg of the trimethoprim component per day) was themostfrequentlyprescribedtreatment,eitherasinitialtherapy oras sequential oraltreatment following parenteral‘induction’ therapywithimipenemwithorwithoutamikacin.Mostpatients receivedantibiotictherapyforaprolongedperiodoftime(median 12weeks;range1–52weeks).
Six patientswere lost tofollow-up, most frequently due to transfertoothermedical institutions.Theoverallmortalityrate was 18.9% (n=7/37) and a trend towards higher mortality in
patients with disseminated nocardiosis was observed (50% vs.
15.4%;p=0.101),withthehighestmortalityrateinpatientswith CNS involvement (75%); all patients with a fatal disseminated infectionhadCNSinvolvement.
Population-basedestimatesofannualratesofhospitalizedPNin Germanybetween2005and2011
From2005to2011,theaverageannualGermanpopulationwas 82.1 million (range 81.1–82.4 million). During this period, an average1673hospitals(95%confidenceinterval(CI)1633–1719) weresubjecttoDRGbillingofmedicalservicesandtheaverage overallnumberofhospitalsinGermanywas2087(95%CI2067– 2108).Therefore,80%ofallhospitalswerecoveredbythestudy analysis.Atotalof125.2millionhospitalizationswereanalyzedfor theperiod2005–2011(anaverage17.9(95%CI17.4–18.3)million hospitalizations per year) over an observation period of 574.4 million person-years. Overall, 266 hospitalizations with PN as either theprimary or secondary diagnosiswere identified. The overallaverageannualage-adjustedhospitalizationratewas0.04 per100000inhabitants:0.05per100000inmalesand0.03per 100000 in females. However,there was considerablevariation withage,withthehighestage-specifichospitalizationrateof0.24/
100000inhabitants among menaged75–84 years(Figure3A).
During the study period, the annual number of PN-associated hospitalizationsremainedrelativelyunchangedbetween2005and 2011forbothmalesandfemales(Figure3B).
Discussion
This study investigated a large cohort of PN patients in European countries, applying a stringent definition of PN. It thereforeprovidesimportantinsightsintotheepidemiologyand clinical appearance and course of this rare infection in high- incomecountries.
Thepivotal role of immunocompromisingcomorbiditiesand treatments as major risk factors for PN has been described previously(Martinezetal.,2008;Mineroetal.,2009;Ambrosioni et al., 2010; Chen et al., 2013; Poonyagariyagorn et al., 2008;
Takiguchi et al., 2017).In the present study cohort, 93% of all patientswere suffering from a relevant comorbidity and 51.2%
Figure2. DistributionofNocardiaspp.(n=37).
Table3
Microbiologicalresults:sourceofclinicalspecimenandpresenceofco-infections (n=43).
Number %
Typeofdiagnosticspecimen
BAL 19/43 44.2%
Sputum 9/43 20.9%
Tracheobronchialaspirate 5/43 11.6%
Lungbiopsy 3/43 7.0%
Brainabscess 3/43 7.0%
Skinandsofttissueabscess 2/43 4.7%
Pleuraleffusion 2/43 4.7%
Pulmonaryco-infectiona 16/43 37.2%
CMV 3/16 18.8%
Staphylococcusaureus 2/16 12.5%
Pseudomonasaeruginosa 2/16 12.5%
MTB 2/16 12.5%
NTM 2/16 12.5%
Aspergillusspp. 2/16 12.5%
Otherb 6/16 37.5%
BAL,bronchoalveolarlavage;CMV,cytomegalovirus;MTB,Mycobacteriumtubercu- losis;NTM,non-tuberculousmycobacteria;SD,standarddeviation.
aPatientscouldhavemorethanone.
b Onepatienteachwithoneofthefollowing:Moraxellaspp.,Escherichiacoli, Stenotrophomonasmaltophilia,Haemophilusparainfluenzae,Pneumocystisjirovecii, Ralstoniapickettii.
werebeingtreatedwithanimmunosuppressivemedication.Most patientswere receiving a combination of various immunosup- pressiveagents,althougheightpatientswerebeingtreatedwith only systemic corticosteroid monotherapy. Single organ trans- plantation(SOT)wasthemostcommonriskfactorinthiscohort.
This confirms the data from mostrecent reports, showing the occurrenceofPNin1.3%to3.5%ofallSOTrecipients(Poonyagar- iyagornetal.,2008;Pelegetal.,2007;Hemmersbach-Milleretal., 2018;Majeedetal.,2018).
PNmayalsooccurinimmunocompetentpatients,mainlythose withapulmonarycomorbidity,especiallyCOPD(Chenetal.,2013;
Huietal.,2003;Chenetal.,2014;Kuraharaetal.,2014;Martinez Tomasetal.,2007).Impairedciliarymotilityandepithelialdamage leadtoimpairedlocalimmunedefence,andsystemiccorticoste- roidsmaypromotenocardialinfection(Elenkov,2004).Overhalfof thepatients(58.1%)had anunderlying pulmonarycomorbidity, most commonly COPD (this was the sole comorbidity in 12 patients).Pulmonarycomorbidity,particularlyCOPDandbronchi- ectasis, has been described as a risk factor for PN in non- immunosuppressedpatients(Woodworthetal.,2017;Steinbrink etal.,2018).
Recent reports addressing radiological abnormalities in PN agreethatthealterationsfoundinPNarediverse,notpathogno- monic,andmaymimicamultitudeofpulmonarydiseases(Chen etal.,2014;Kuraharaetal.,2014;Blackmonetal.,2011;Tsujimoto et al., 2012). Despite the non-discriminative and non-specific radiographicpictureofPN,thisstudyisnovelindemonstrating thatPNpredominantlyaffectstherightupperlobe,evenafterthe exclusionofco-infectedpatientsfromtheanalyses.However,this contradictsthefindingsofpreviousstudies,whichhaveshownno
particular distribution of radiographic abnormalities or even a lowerlobepredominance(Chenetal.,2014;Blackmonetal.,2011;
Tsujimoto et al., 2012). A possible explanation could be the stringentdefinitionofPNappliedinthepresentstudy,ensuring theinclusionof‘true’PNonly.Furthermore,theaerobicnatureof Nocardiaspp.andthefactthattherightupperlobeisregardedas one of the best ventilated areas of the lungs may potentially explainthegreaterupperlobeinvolvement.
Duetothenon-specificclinicalandradiologicalpresentation,a delaybetweentheonsetofsymptomsandadefinitivePNdiagnosis iscommon,andPNisinitiallyoftenmisdiagnosed.Thediagnostic delay of 14daysin thestudy cohort is in line withthe recent literature,reportingmeandiagnosticintervalsof13.6to42days (Poonyagariyagornetal.,2008;MartinezTomasetal.,2007).The diagnosticdelaymayalsobeattributedtothedelayedgrowthin culture,whichmaytakeupto3weeks(Ambrosionietal.,2010;
Ashdown,1990).Therefore, the microbiologylaboratory should always becontactedwhena nocardial infectionissuspectedto ensure that the corresponding specimens are incubated for a Table4
Resultsofsusceptibilitytesting.a
Susceptibility
TMP–SMX Amikacin Imipenem Benzylpenicillin Piperacillin
Nocardiaspp.(n=4) 100%(4) 100%(4) 100%(4) 75%(3) 100%(4)
Nocardiaabscessus(n=4) 75%(3) 100%(4) 50%(2) 75%(3) 75%(3)
Nocardiaasteroides(n=3) 100%(3) 100%(3) 66.7%(2) 100%(3) 100%(3)
Nocardiafarcinica(n=3) 100%(3) 100%(3) 33.3%(1) 100%(3) 100%(3)
Nocardiacyriacigeorgica(n=2) 100%(2) 100%(2) 100%(2) 100%(2) 100%(2)
Nocardiaveterana(n=3) 100%(3) 100%(3) 100%(3) 66.7%(2) 100%(3)
Nocardiaotitidiscaviarum(n=1) 100%(1) 100%(1) 100%(1) 100%(1) 100%(1)
Total(n=20) 95%(19) 100%(20) 75%(15) 85%(17) 95%(19)
TMP–SMX:trimethoprim–sulfamethoxazole.
aBreakpointsforresistancearethoseoftheClinicalandLaboratoryStandardsInstitute(CLSI),versionM24A2(2011).
Table5
Antibiotictreatmentof43patientswithpulmonarynocardiosis.
Number %
Antibioticsa
TMP–SMX 22 51.2%
Carbapenem 18 41.9%
Amikacin 12 27.9%
Cephalosporin 8 18.6%
Penicillin 8 18.6%
Quinolone 5 11.6%
Tetracycline 4 9.3%
Linezolid 4 9.3%
Macrolide 3 7.0%
Otherb 8 18.6%
MeanSD Median(range) Durationoftreatment(weeks)(n=37) 15.614.9 12(1–52) SD,standarddeviation;TMP–SMX,trimethoprim–sulfamethoxazole.
aPatientscouldhavemorethanone.
b Standardanti-tuberculoustreatment(n=4),glycopeptide(n=2),sulfadiazine (n=1),clindamycin(n=1).
Figure3. Population-basedestimatesofannualincidenceratesofhospitalized caseswithpulmonarynocardiosis:(A)averageannualage-specifichospitalization rate;(B)annualage-adjustedincidenceratesofhospitalizedpulmonarynocardiosis inGermanyfrom2005to2011.
prolonged period of time and to enable early application of moleculartechniquesforthedetectionoftheorganism(Ambro- sionietal.,2010).
In this cohort, 18.4% of all evaluable patients had a disseminatedinfection,mainlywithCNSinvolvement.However, evaluablecerebralCTscanswereavailableforonly12patients;
therefore, the true incidence of this disseminated course is possibly underestimated. An active search for dissemination, includingcerebralCTscanormagneticresonanceimagingand ophthalmological investigations (uveitis), is of particular importance, because dissemination is directly linked to the outcomeandtheselectionofdefinitiveantibiotictherapy(CNS penetration)(Ambrosionietal.,2010;CortiandVillafane-Fioti, 2003).
The most commonly used antibiotics were TMP–SMX, imipenem, amikacin, and ceftriaxone, which are all generally recommendedas initial treatmentfor PN (Brown-Elliott et al., 2006;Wilson,2012).Inmostpatients,atreatmentdurationof6 months is considered sufficient, whereas in cases with CNS involvementorinseverelyimmunosuppressedcases,evenmore prolongedtreatmentmaybenecessary(Ambrosionietal.,2010;
Corti and Villafane-Fioti, 2003). However, some authors have recommendedshorter treatmentcourses (Tripodiet al., 2011).
Theresults ofthe presentstudy,witha highclinical curerate (overall>80%andalmost90%innon-disseminatedcases)anda mediantreatmentdurationof12weeksmayprovideadditional support for futureshortening of antibiotic therapy in selected patients.
Somespeciespresentedclinicallyrelevantresistancethatmay leadtoinadequatetreatmentandtreatmentfailure.N.farcinicaand N.asteroides,thetwomostfrequentspeciesinthiscohort,were susceptible to imipenem in only 33.3% and 66.7% of cases, respectively,whereassusceptibilitytoallothertestedantibiotics was100%.Ofnote,amikacinwastheonlytesteddrugwithproven in vitro efficacy against all Nocardia abscessus isolates in this cohort. The study findings add to the available data on susceptibility testing and demonstrate increasing resistance to commonlyrecommendedantibiotics(Mineroetal.,2009;Uhde etal.,2010;Lebeauxetal.,2018;Schlabergetal.,2014;Valdezate etal.,2017;McTaggartetal.,2015).Inaddition,thedataareinline withthoseofarecentstudyfromFrance,showingthatN.farcinica is the most common species in human infection in Western Europe,witha distinct patternof susceptibility(Lebeaux etal., 2018).Earlyspeciesidentificationusing16sRNAPCRsequencingor multilocussequencingtypingandsusceptibilitytestingmayhelp toimproveantibiotictreatment,althoughtheclinicalimpactofthe observedresistanceremainsunclear.
In contrasttoearlier case series and recent epidemiological studiesbasedonlaboratorytestingonly,amajorstrengthofthe presentstudyistheclearandrestrictivedefinitionofPN,making theunintendedinclusionofcaseswithcolonizationimprobable.
Although Nocardia spp. are neither commensals of the normal human flora nor common laboratory contaminants, the sole detection of Nocardia spp. in respiratory samples, especially sputum, in the absence of clinical or radiological signs and symptoms,doesnotnecessarilyrepresentaninfectioninprogress.
Anotherstrengthof this studyisthefactthat patientsfroman unselectedcohortwereincluded(allpatientsfromtheparticipat- ing institutions) rather than searching for cases in selected populationssuchasHIV-positivepatientsorSOTandbonemarrow transplant recipients. The data therefore reflect the complete clinical,microbiological, and radiographic spectrum of PN that may be encountered in general medical departments and pulmonologydivisions.
This was a retrospective evaluationwithall the limitations that come with a retrospective study design, especially the
problemofmissingdata.TheidentificationofNocardiaspp.and susceptibilitytestingwereperformedonlyuponspecialrequest.
Therefore, the species distribution may not reflect the real importanceof eachspecies in PN.Thenumbersof patientsper countryvariedconsiderably,makingcomparisonbetweencoun- tries impossible. However, since all four countries provide comparable standards of medical care in terms of diagnostic facilities, availability of therapiesand medical procedures, and prevalenceofcomorbiditiesconsideredastypicalriskfactorsfor nocardiosis, it is believed that the pooled data may be representative of Western European countries, although not generallytransferabletoothercountries.
The population-based estimate of the hospitalization rate (0.04/100000 inhabitants)mayserveasasurrogate forannual incidence and is considerably lower than rates reported previously: 0.39–0.55/100000 inhabitants in Spain (0.20 if extrapolated tothewholepopulation ofSpain)and 0.33–0.87/
100000 inhabitants in Canada (Minero et al., 2009; Tremblay etal.,2011).Theremaybeimportantreasonsforthisdifference.
The present study focused on pulmonary nocardiosis only, whereastheothercohortsincludednocardiosisindependentof thesiteofinfection.Furthermore,inbothpreviousstudies,cases wereidentifiedbymicrobiologicalproofofNocardiaspp.,without differentiatingbetweeninfectionandcolonization.Incontrast,the present study estimates are based on the coding of hospital dischargediagnosesforbilling(ICD-10codes),makingcoloniza- tionunlikely,becausediagnosticandtherapeuticeffortshavetobe justified for reimbursement and medical charts are frequently reviewed by health insurance companies. However, they are consideredtohavehighspecificity,butonlymoderatesensitivity, thus being proneto an underestimation of disease prevalence (O’Malleyetal.,2005).
Thestudyestimatesarethefirsttobebasedonarepresentative nationwide population-based analysis, including 80% of all hospitalsacrossGermanyand>125millionhospitalizationsover a7-yearperiod.Therefore,apossiblebiasduetodifferencesinPN- associatedhospitalizationsinruralandurbanareas,mainlydueto potential regionaldifferences in health care utilization, can be excluded. However,itwas not possibletoaccount forreadmis- sions.Finally,theresultsapplytohospitalizedpopulationsonly.PN isoftenachronicinfectionandusuallyrequireslong-termfollow- up care in the outpatient setting, where disease prevalence is presumablydifferent.Therefore,thedataarelikelytounderesti- mate the overall burden of PN. However, asavailable data are limited,theresultsarethebestcurrentlyavailablesurrogatefor theepidemiologicaltrendsofPNacrossalargeWesternEuropean country(Germany).
IncontrasttotheCanadianandtheSpanishcohorts,notrend towardsanincreasingincidenceovertimewasobservedinthis study. This previously described trend was explained by an increasingnumberofimmunocompromisedandimmunosenes- cent patients in developed countries, improved laboratory detection techniques, and an increased awareness of PN in patients at risk during the observation periods (Minero et al., 2009;Tremblayetal.,2011;Agterofetal.,2007).Thediscrepancy between the present study and previous studies might be explained by differences in the observation period. Minero etal.evaluatedincidencesbetween1995and2006andTremblay etal.between1988and2008,whereasthisstudyanalyzedthe periodbetween 2005and 2011 (Mineroetal., 2009;Tremblay et al., 2011). Therefore, immunosuppressive treatment with biologicalmedicines,whichwasintroduced intheearly2000s, as well as genotypic methods for species identification were already in widespread use during the study period and, furthermore, the number of SOT performed annually did not increasesignificantlyduringthestudyperiod,incontrasttothe
previousstudies(https://www.organdonor.gov/statistics-stories/
statistics/data.html).
Theepidemiologicaldataconfirmedthehighestincidencerate ofPNinelderlymales,whichhaspreviouslybeenassumedfrom case series and case reports (Figure 1A) (Minero et al., 2009;
Kurahara et al., 2014). For the first time, a discrete female predominancewasobservedamongyoungeradults.Thismightbe attributable tothe female predominance in most autoimmune connectivetissuedisorderswithsubsequentimmunosuppressive treatment.However,thedataarenotsufficienttodrawdefinitive conclusionsandfurtherstudiesareneededtoestablishthisasa distinctphenotypeofPNpatients.
In conclusion, this study shows that PN remains a rare pulmonaryinfectionthatprimarilyaffectsimmunocompromised patients.Between2005and2011,thePN-associatedhospitaliza- tionratesremainedunchanged.Theclinicalappearanceisnon- specific; heightened awareness and clinical suspicion are of particular importance in patients at risk. Dissemination to extrapulmonary organs, especially tothe CNS, is frequent and associated with a worse outcome. Cerebral imaging should be prompted in all patients once the diagnosis of pulmonary nocardiosis is settled. Different Nocardia spp. may exhibit distinctivepatternsofresistance,makingidentificationofNocardia spp.beyondjustrecognitionofthegenusindispensable.Therefore, timely prescription of empiric combination therapy may be beneficialuntiltheresultsofsusceptibilitytestingareavailable.
Acknowledgement
Wewouldliketothankallmembersandcontributorstothe OrphanedPulmonaryInfectionNetwork(OPINION)StudyGroup:
Prof. Dr Torsten Bauer (Berlin, Germany), Dr Frank Bergmann (Berlin, Germany), Prof. Dr Dieter Buchheidt (Mannheim, Germany), DrAndres de Roux (Berlin, Germany), Dr SaraDroz (Bern, Switzerland), Dr Holger Flick (Graz, Austria), Dr Hilte Geerdes-Fenge(Rostock,Germany),Prof.DrAndreaGrisold(Graz, Austria), PD Dr Martin Kolditz (Dresden, Germany), Prof. Dr Stephen Leib (Bern, Switzerland), PD Dr Philipp M. Lepper (Homburg/Saar,Germany),DrSebastianR.Ott(Bern,Switzerland), Prof. Dr Mathias Pletz (Jena, Germany), Dr Elisabeth Presterl (Vienna,Austria), Dr BrunoRobibaro (Vienna,Austria), Prof.Dr GernotRohde(Maastricht,TheNetherlands),DrDirkSchürmann (Berlin,Germany),DrThomasValentin (Graz,Austria),DrKlaus Vander(Graz,Austria),DrSilvan Vesenbeckh(Berlin,Germany).
TheOPINIONstudygroupisaprojectofthePaul-Ehrlich-Society (PEG),Germany.M.W.P.wassupportedbytheGermanMinistryof EducationandResearch(BMBF),grantnumber01KI1501.
Fundingsource
Thiswork was supportedby thePaul-Ehrlich-Society(PEG), Germany.TheOPINIONStudyGroupisaprojectofthePEG.M.W.P.
wassupportedbytheGermanMinistryofEducationandResearch (BMBF), grant number 01KI1501. None of the authors has a competinginteresttodeclare.
Ethicalapproval
The study was approved by the ethics committees at the participating sites. Informed consent was judged not to be necessaryduetotheretrospectivenatureofthestudy.
Conflictofinterest
Noneoftheauthorshasaconflictofinteresttodisclose.
References
AgterofMJ,vanderBruggenT,TersmetteM,terBorgEJ,vandenBoschJM,Biesma DH.Nocardiosis:acaseseriesandaminireviewofclinicalandmicrobiological features.NethJMed2007;65(6):199–202.
AmbrosioniJ,LewD,GarbinoJ.Nocardiosis:updatedclinicalreviewandexperience atatertiarycenter.Infection2010;38(2):89–97.
AshdownLR.AnimprovedscreeningtechniqueforisolationofNocardiaspecies fromsputumspecimens.Pathology1990;22(3):157–61.
BeamanBL,BeamanL.Nocardiaspecies:host-parasiterelationships.ClinMicrobiol Rev1994;7(2):213–64.
BeamanBL,BurnsideJ,EdwardsB,CauseyW.NocardialinfectionsintheUnited States:1972-1974.JInfectDis1976;134(3):286–9.
BlackmonKN,RavenelJG,GomezJM,CiolinoJ,WrayDW.Pulmonarynocardiosis:
computedtomographyfeaturesatdiagnosis.JThoracImaging2011;26(3):224–
9.
BoironP,etal.ReviewofnocardialinfectionsinFrance1987to1990.EurJClin MicrobiolInfectDis1992;11(8):709–14.
Brown-ElliottBA,BrownJM,ConvillePS,WallaceJr.RJ.Clinicalandlaboratory features ofthe Nocardia spp.based on currentmolecular taxonomy.Clin MicrobiolRev2006;19(2):259–82.
ChenYC,LeeCH,ChienCC,ChaoTL,LinWC,LiuJW.Pulmonarynocardiosisin southernTaiwan.JMicrobiolImmunolInfect2013;46(6):441–7.
ChenJ,ZhouH,XuP,ZhangP,MaS,ZhouJ.Clinicalandradiographiccharacteristics of pulmonary nocardiosis:clues to earlierdiagnosis. PLoS One 2014;9(3) e90724.
CortiME,Villafane-FiotiMF.Nocardiosis:areview.IntJInfectDis2003;7(4):243–
50.
DeemRL,DoughtyFA,BeamanBL.ImmunologicallyspecificdirectTlymphocyte- mediatedkillingofNocardiaasteroides.JImmunol1983;130(5):2401–6.
Elenkov IJ. Glucocorticoids and the Th1/Th2 balance. Ann N Y Acad Sci 2004;1024:138–46.
FiliceGA,NiewoehnerDE. Contributionofneutrophilsandcell-mediatedimmunityto controlofNocardiaasteroidesinmurinelungs.JInfectDis1987;156(1):113–21.
GermanFederalStatisticalOffice.Facts&Figures—State&Society—Federal StatisticalOffice(Destatis). 2013[cited25February2013];Availablefrom:
www.destatis.de/EN/FactsFigures/SocietyState/SocietyState.html.
HardakE,YiglaM,BergerG,SprecherH,OrenI.Clinicalspectrumandoutcomeof Nocardiainfection:experienceof15-yearperiodfromasingletertiarymedical center.AmJMedSci2012;343(4):286–90.
Hemmersbach-MillerM,StoutJE,WoodworthMH,CoxGM,SaulloJL.Nocardia infectionsinthetransplantedhost.TransplInfectDis2018;20(4)e12902.
HuiCH,AuVW,RowlandK,SlavotinekJP,GordonDL.Pulmonarynocardiosis re-visited:experienceof35patientsatdiagnosis.RespirMed2003;97 (6):709–17.
KuraharaY,TachibanaK,TsuyuguchiK,AkiraM,SuzukiK,HayashiS.Pulmonary nocardiosis:aclinicalanalysisof59cases.RespirInvestig2014;52(3):160–6.
LebeauxD,BergeronE,BerthetJ,Djadi-PratJ,MouniéeD,BoironP,etal.Antibiotic susceptibility testing and species identification of Nocardia isolates: a retrospectiveanalysisofdatafromaFrenchexpertlaboratory,2010-2015.Clin MicrobiolInfect2018;,doi:http://dx.doi.org/10.1111/tid.12904.
MajeedA,BeattyN,IftikharA,MushtaqA,FisherJ,GaynorP,etal.A20-year experiencewithnocardiosisinsolidorgantransplant(SOT)recipientsinthe SouthwesternUnitedStates:asingle-centerstudy.TransplInfectDis2018;20(4) e12904.
MartínezTomásR,MenéndezVillanuevaR,ReyesCalzadaS,SantosDurantezM, VallésTarazonaJM,ModestoAlapontM,etal.Pulmonarynocardiosis:risk factorsandoutcomes.Respirology2007;12(3):394–400.
MartinezR,ReyesS,MenendezR.Pulmonarynocardiosis:riskfactors,clinical features,diagnosisandprognosis.CurrOpinPulmMed2008;14(3):219–27.
McNeil MM, Brown JM. The medically important aerobic actinomycetes:
epidemiologyandmicrobiology.ClinMicrobiolRev1994;7(3):357–417.
McTaggartLR,DoucetJ,WitkowskaM,RichardsonSE.Antimicrobialsusceptibility amongclinicalNocardiaspeciesidentifiedbymultilocussequenceanalysis.
AntimicrobAgentsChemother2015;59(1):269–75.
MineroMV,MarínM,CercenadoE,RabadánPM,BouzaE,MuñozP.Nocardiosisat theturnofthecentury.Medicine(Baltimore)2009;88(4):250–61.
O’MalleyKJ,CookKF, PriceMD,WildesKR,HurdleJF,Ashton CM.Measuring diagnoses:ICDcodeaccuracy.HealthServRes2005;40(5Pt2):1620–39.
PelegAY,HusainS,QureshiZA,SilveiraFP,SarumiM,ShuttKA,etal.Riskfactors, clinical characteristics, and outcome of Nocardia infection in organ transplant recipients: a matched case-control study. Clin Infect Dis 2007;44(10):1307–14.
PoonyagariyagornHK,GershmanA,AveryR,MinaiO,BlazeyH,AsamotoK,etal.
ChallengesinthediagnosisandmanagementofNocardiainfectionsinlung transplantrecipients.TransplInfectDis2008;10(6):403–8.
SchlabergR,FisherMA,HansonKE.SusceptibilityprofilesofNocardiaisolatesbased oncurrenttaxonomy.AntimicrobAgentsChemother2014;58(2):795–800.
SteinbrinkJ,LeavensJ,KauffmanCA,MiceliMH.Manifestationsandoutcomesof nocardiainfections:comparisonofimmunocompromisedandnonimmuno- compromisedadultpatients.Medicine(Baltimore)2018;97(40)e12436.
TakiguchiY,IshizakiS,KobayashiT,SatoS,HashimotoY,SurugaY.Pulmonary Nocardiosis:aclinicalanalysisof30cases.InternMed2017;56(12):1485–90.
TremblayJ,ThibertL,AlarieI,ValiquetteL,PépinJ.NocardiosisinQuebec,Canada, 1988-2008.ClinMicrobiolInfect2011;17(5):690–6.
TripodiMF,Durante-MangoniE,FortunatoR,CuccurulloS,MikamiY,FarinaC,etal.
In vitro activity of multiple antibiotic combinations against Nocardia:
relationshipwithashort-termtreatmentstrategyinhearttransplantrecipients withpulmonarynocardiosis.TransplInfectDis2011;13(4):335–43.
TsujimotoN,SarayaT,KikuchiK,TakataS,KuriharaY,HiraokaS,etal.High- resolutionCTfindingsofpatientswithpulmonarynocardiosis.JThoracDis 2012;4(6):577–82.
UhdeKB,PathakS,McCullumJrI,Jannat-KhahDP,ShadomySV,DykewiczCA,etal.
Antimicrobial-resistantnocardiaisolates:UnitedStates,1995-2004.ClinInfect Dis2010;51(12):1445–8.
ValdezateS,GarridoN,CarrascoG,Medina-PascualMJ,VillalónP,NavarroAM,etal.
EpidemiologyandsusceptibilitytoantimicrobialagentsofthemainNocardia speciesinSpain.JAntimicrobChemother2017;72(3):754–61.
WilsonJW.Nocardiosis:updatesandclinicaloverview.MayoClinProc2012;87 (4):403–7.
WoodworthMH, SaulloJL, LantosPM, CoxGM,StoutJE.Increasing Nocardia incidenceassociatedwithbronchiectasisatatertiarycarecenter.AnnAm ThoracSoc2017;14(3):347–54.
