Analysis Of The Etiologies Of Fever Of Unknown Origin With Pulmonary Infiltratesand The Value Of Xpert MTB/RIF Assay For Diagnosis Of Suspected Pulmonary Tuberculosis

Analysis of the Etiologies of Fever of Unknown Origin with Pulmonary InfiltratesBackground and Objective: Fever of unknown origin(FUO)with pulmonary infiltratesis a common symptom in respiratory medicine,can be caused by various causes of infectious and non-infectious diseases,such as infection,neoplasm,connectivetissuediseaseandmiscellaneous.Therefore,it is more important to identify the cause for FUO patientswith pulmonary infiltratesbefore the correct treatment strategy.However,following factors make the FUO with pulmonary infiltrates diseases difficult to be diagnosed: atypical clinical manifestations,poor effective detection methods,the different imaging manifestations in the same disease,the similar imaging performance in different diseases.Forexample,non-infectious diseases,such ashematologic diseases,lung cancer,pulmonary embolism,connective tissue disease,cryptogenic organizing pneumonia,eosinophilic pneumonia.et can have image of lobar or miliary infiltrates,cavitas andother performance.As all of the diseases mentioned above might have the same imaging performance aspneumonia does,misdiagnosing noninflammatory disorders aspneumonia is not rare in clinical setting.As reported,the clinical misdiagnosis rate was up to 17% to 47%[1-2],which will be certainly involving in overdiagnosis or treatment delay.The overuse of antibiotics increase the risks for antibiotic resistance while under-diagnosed will delay the treatment and affect the patient’s outcome.So,timely updating and mastering the disease spectrum of FUO in this area,summarizing experience and lessons,will certainlyhelp clinicians more fully understand FUO with pulmonary infiltrates.This study is toexplore the etiologies and its clinical characteristics of FUOwith pulmonary infiltrates in our hospital in recent years.Methods:A retrospective analysis was made in 104 patients that meeting thecriteria of FUO with pulmonary infiltrates,who hospitalized in respiratory medicine in the First Affiliated Hospital of Guangzhou Medical University during July 2012 to December2016.Each patient’s clinical data was recorded.The final diagnosis is confirmed by the diagnostic criteria or guidelines for each disease[3-4].Based on the discharge diagnosis they were divided intofive groups: infectious diseases;connective tissue disease;neoplasm;miscellaneousand nodiagnosis.The basic informations of each group were compared and analyzed,such asgender,age,fever duration(the duration ofillness before hospitalization),diagnostic methods.Then we summarize the regularity.Results:Of the 104 patients,93 cases(89.4%)were discharged with definitive diagnosis.Among them,48 cases(46.2%)were caused by infectious diseases,29 cases(27.9%)were diagnosed as connective tissue disease,14 cases(13.5%)were suffered from neoplasm,2 cases(1.9%)were miscellaneous,andthe last 11 cases(10.6%)were still unknown after hospitalization.Infectious disease isthe main cause of FUOwith pulmonary infiltrates.In infectious disease,15 cases turned to be tuberculosis,accounted for the largest proportion of 31.3%(15/48),followed by fungal infection,accounted for27.1%(13/48),including Aspergillus infection(4 cases),Penicillium marneffei(3 cases).In addition,there are also 3 cases of non-tuberculous mycobacteria(NTM),2 cases of narcissus,and 1 case of brucellosis.Virus infection accounted for 8.3%(4/48)(including new type of human avian influenza virus in 3 cases and EB virus in 1 case).In connective tissue disease,vasculitis was the most common disease,accounted for32.1%(9/29),followed by adult still’s disease and cryptogenicorganizing pneumonia(COP),accounted for 13.8%(4/29),respectively.In neoplasm,lymphoma and lung cancer were the most common diseases,accounted for 42.9%(6/14)and 35.7%(5/14),respectively.In addition,there were two Langerhans cell histiocytosisof miscellaneous.Themedian duration of feverin neoplasmgroupwas 31.5 days(range,15-360 days),with25th/75 th percentilesbeing(27.25-165.0)days.Themedian duration of feverininfectious disease was 28 days(range,14-180 days),with 25th/75 th percentilesbeing(21-60)days.Themedian duration of feverin connective tissue disease was 30 days(range,21-360 days),with 25th/75 th percentilesbeing(24-75)days.The duration of fever in infectious groupwas shorterthan that in neoplasmsgroup andconnective tissue disease group.There were statistical differences between each groups(P=0.03).44.2%(46/104)fever durationwere within 30 days,of which 67.4%(31/46)were infectious diseases.Among 93 patients with confirmed diagnosis,lung tissue or bone marrow biopsy is the main method,performing 47.3%(44/93).The proportion of biopsy in infectious disease,connective tissue disease,neoplasm and miscellaneous were 35.4%(17/48),64.3%(9/14),55.2%(16/29)and 100%(2/2),respectively.11.8%(11/93)were diagnosed by CT/PET-CT/MRI(lung cancer were diagnosed by PET-CT in 4 case(4/7)).Twenty-one out of 48 infectious diseases(43.8%)were diagnosed by pathogenic detection(culture in 16 cases,nucleic acid amplification tests in 3 cases,sputum smear in2 cases).In the imaging of 93 patients with confirmed diagnosis,there were 49.5%(46/93)demonstrating lobar or miliary infiltrates,followed by multiple nodular shadows,accounted for 16.1%(15/93),effusion and nodular shadowsa were 10.8%(10/93).Besides,18.3%(17/93)caseswere combined with cavities.34.4%(32/93)were combinedwith pleural effusion,and the proportion was.28.0%(26/93)were lung and/or mediastinal lymphadenopathy.Patients with non-infectious diseaseswere acompanying with infection at admission,accounting for 35.6%(16/45),due to fever was still existed after infection was under controled,and all of them were diagnosed by pathology eventually.Conclusion:Infectious diseases,connective tissue disease and neoplasm are still the three major causesof FUO with pulmonary infiltrates.Infectious diseases,especially TB,continue to be the leading etiology of FUO with pulmonary infiltrates,but fungal infection should not be ignored.Microbiological sputum smear combined with molecular examination has good value for the diagnosis of infection diseases.Thoughinvasive operations such as lung biopsy may have some risk to patients,it is of great significance fordiagnosisof FUO patients with pulmonary infiltrates,especially for the ones with neoplasm.Specific imaging performance is useful for the diagnosis of FUO patients with pulmonary infiltrates.By combining a variety of diagnostic techniques,most of the FUO patients with pulmonary infiltrates eventually can be clearly diagnosed.Evaluation of the Xpert MTB/RIF assay for diagnosis of suspected pulmonary tuberculosisBackground and Objective: As the most common cause of FUO in the present study,tuberculosis is also common in FUO patients with pulmonary infiltrates.However,following factors make TB difficult to be quickly and accuratelydiagnosed: atypical clinical manifestations,negative outcomes of conventional laboratory methods such as sputum smear and PPD.Therefore,rapid andimproved TB diagnosis is urgently needed.The application of T cell spots detected in tuberculosis infection(T-SPOT.TB),Tuberculosisdeoxyribonucleic acid(TB-DNA),and Xpert MTB/RIF(Cepheid)assayhave been developed.There was little information regarding the performance characteristics betweenthem.And very little information regarding the performance characteristics of Xpert MTB/RIF assay is available.The purpose of this study was to evaluate the performance of Xpert MTB/RIF assay compared toconventional sputum smear,TB-DNA and T-SPOT.TBin diagnosingpulmonary tuberculosis from the aspects of detection rate,sensitivity,and detection time.Methods:A total of 283 suspected pulmonary tuberculosis(with gradiographic imaging features of TB)were enrolledwho wereadmitted from First Affiliated Hospital of Guangzhou Medical University during August 2015 to March 2017.Patients who were taking TB treatment formore than 2 weeks prior to the initiation of the studywere excluded.Basic clinical data of all patientswere recorded.Two or threesputum specimenswere collectedand tested withsmear microscopy in each,microbacterial culture(using liquid or solid media),TB-DNA,and Xpert MTB/RIF assay.283 cases were examined directly by sputum smearand Xpert MTB/RIF assay,whilst 260 cases were detected by sputum smearand TB-DNA,186 cases had T-SPOT.TB testing,178 cases were tested with sputum smear,Xpert MTB/RIF,TB-DNA and T-SPOT.TB in parallel.The diagnosis of TB is based on 《 Diagnostic criteria for pulmonary tuberculosis 》(ws288-2008).Each patient was allocated to one of two diagnostic categories.1) Confirmed TB: a clinical presentation compatible with TB with at least one sputum sample smear positive or culture positive for M.Tuberculosisor or pathological biopsy proved to be tuberculosis;2)Non-confirmed TB:no evidence of TB based on smear microscopy,culture,and Lung biopsy to support the diagnosis of TB,diseases that were not proven to be TB.The Confirmed TB group was divided into smear-positive tuberculosis group and smear-negative tuberculosis group according to smear status.Of the 283 patients,81 were categorized as confirmed pulmonary TB(33 casesin smear-positive pulmonary tuberculosis and 48 cases in smear-negative pulmonary tuberculosis),202 were not proven to be TB as non-confirmed TB group.The sensitivity and specificityof smear microscopy,Xpert MTB/RIF assay,TB-DNAand T-SPOT.TB in detecting MTB were analyzed.The positive rates of smear microscopy,and Xpert MTB/RIF for M.tb were12.7%(36/283)and 25.4%,(72/283),respectively.Xpert MTB/RIF assay shown positive in 38(46.9%,38/81)comfiremed TB withsmear negative..Smear-positive/culture-positive or pathology consistent with tuberculosis was used as a reference standard for confirmed TB.Sensitivity calculations of Xpert MTB/RIFtake into account only patients with confrmed TB.The sensitivity,specificity,as well as the positive and negative predictive value of Xpert MTB/RIF assay were 87.7%(71/81),99.5%(201/202),98.6%(71/72)and95.3%(201/211),respectively.The sensitivityand specificity ofsmear microscopy examination was 40.7%(33/81)and 98.5%(199/202),repsectively,PPV and NPV was91.7%(33/36)and 80.6%(199/247),repsectively.T-SPOT.TBresult demonstrated sensitivity of75.5%(37/49),specificity of 70.1%(96/137),positivepredictive value(PPV)of 47.4%(37/78),and negative predictive value(NPV)of 88.9%(96/108).The sensitivity,specificity,PPV and NPV of TB-DNA in detecting MTB was60.3%(41/68),97.4%(187/192),89.1%(41/46)and 87.4%(187/214),respectively.When the specimen stratified by smear microscopyresult,the sensitivity of Xpert MTB/RIF,T-SPOT.TB and TB-DNA was 100%(14/14),92.9%(13/14)and 85.7%(12/14)for smear-positive TB,and77.4%(24/31),71.0%(22/31),51.6%(16/31)for smear-negative TB,respectively.The sensitivity of smear-positive pulmonary tuberculosis group was higher than that of smear-negative tuberculosis group.Although Xpert MTB/RIF assayremained a higher sensitivity for smear-negative pulmonary tuberculosis,there were still about 20.0% false-negative resultsinsmear-negative pulmonary tuberculosis.By observing detection indicators of the sputum smear,T-SPOT.TB and TB-DNA,we found that smear microscopy had a good rule-in value for TB;T-SPOT.TBcould have beenruled out as non-TB prior torule-in value;TB-DNAcould have beenrule-in value prior toruled out as non-TB.In order to play the advantages of each method,we will carry out several combined strategies,hoping to improve higher diagnostic value relative to individual Xpert MTB/RIF.Adjunctive diagnosticstrategies are as follows: a)Smear microscopyfollowed by Xpert MTB/RIF(performed if smear negative);b)T-SPOT.TB followed by Xpert MTB/RIF(performed if T-SPOT.TB positive);c)Smear microscopyfollowed T-SPOT.TB(performed if smear negative)and Xpert MTB/RIF(performed if T-SPOT.TB positive);d)TB-DNA followed by Xpert MTB/RIF(performed if TB-DNA negative);e)Smear microscopyfollowed TB-DNA(performed if smear negative)and Xpert MTB/RIF(performed if TB-DNA negative).The receiver operating characteristic curve(ROC)was plotted and the area under the curve(AUC)was evaluated to assess the diagnostic value of the above-mentioned adjunctive diagnosticstrategies.The higher the AUC is,the higher the diagnostic value it has.Results are as follows: Individual Xpert MTB/RIF detection has the highest AUC(0.918),Smear microscopyfollowed by Xpert MTB/RIF is the second,the AUC is 0.915.The AUC of(b)T-SPOT.TB followed by Xpert MTB/RIF,(c)Smear microscopyfollowed T-SPOT.TB and Xpert MTB/RIF,(d)TB-DNA followed by Xpert MTB/RIF,(e)Smear microscopyfollowed TB-DNA and Xpert MTB/RIF are 0.852 、 0.859 、 0.914 、 0.911,respectively.The sensitivity of(b)T-SPOT.TB followed by Xpert MTB/RIF and(d)TB-DNA followed by Xpert MTB/RIF for smear-negative are 61.3%,80.6%,respectively.The specificity of them is 99.2% and 96.2%,respectively.Compared to individual Xpert MTB/RIF assay,these twoadjunct tests did not improve the detection rate of smear-negative pulmonary tuberculosis patients.Conclusion: 1)Compariedto sputum smear,T-SPOT.TB and TB-DNA,the sensitivity and specificity of Xpert MTB/RIF are relatively high in the diagnosis of tuberculosis.Xpert MTB/RIFcan greatly improve the detection rate of tuberculosis,especially for part of smear-negative tuberculosis.It is easy to operate and rapid to provide results within 2.5h;2)No adjunct test was significantly better than the values of individual Xpert MTB/RIF assay in the diagnosis of TB;3)Due to possible of NTM in sputum smear-positive,Xpert MTB/RIF assay performed directly on suspected pulmonary tuberculosis patients is sufficient to detect pulmonary tuberculosis early and shorten the time to distinguish.TB with NTM.