Rapidly Discriminating Culture-negative Urine Specimens From Patients With Suspected Urinary Tract Infections By UF-5000

Background and objectiveUrinary tract infections(UTIs)are one of the most common bacterial infections in community populations and hospitalized patients,and quantitative urine bacterial culture is the "gold standard" for diagnosing UTIs.However,the prevalence of culture-negative specimens is very high;the rate of negative results can be as high as 60%.Additionally,the bacterial culture procedure is time-consuming(at least 24 h)and requires a significant workload from clinical laboratory staff.Finally,clinical laboratory cultures provide insufficient information about the pathogen;thus,many physicians empirically prescribe antibiotics,the effects of which may depend on skill and experience of the physician.Meanwhile,the overuse of antibiotics may lead to antimicrobial resistance in UTIs.Over the decade,automated urine particle analyzers,such as UF-1000i(Sysmex Corporation,Kobe,Japan),have been evaluated as alternatives for screening culture-negative specimens,and these devices have been widely used to evaluate UTI patients.Now,an upgraded analyzer of the UF-Series,UF-5000,has launched,and this device is based on fluorescence flow cytometry technology,using a new blue semi-conductor laser at 488nm wave-length.UF-5000 can rapidly count,recognize and classify urine particles using forward scatter light,side scatter light,side fluorescent light and depolarized side scattered light,in two chambers,surface(SFch)and core(CRch).UF-5000 is an analyser of great interest to detect urine bacterium that can provide new BactInfo-flag information.This information use fluorescent light signals to differentiate between Gram positive and Gram negative,according to different fluorescent-stain behavior on cell wall structures of bacteria.Based on scatter distribution,the Gram Negative?Flag can be inferred that Gram negative bacteria are present.So it can provide important clues for guiding the empirical treatment of UTI patients with antibiotics.A previous study showed that UF-5000 exhibits good performance at detecting urine particles,but there have not been studies investigating its performance at detecting UTIs.This preliminary study aimed to evaluate the performance of new automated UF-5000 analyzer as a rapid method to screen culture-negative specimenns in patients with suspected UTIs and to investigate the utility of the new Bactinfo-flag information for bacterial Gram-stain assessments.Materials&MethodsEthical approval&PatientsThis study was approved by the Research Ethics Committee of the First Affiliated Hospital,School of Medicine,Zhejiang University(reference number 2016-502).Written informed consent was obtained from all patients prior to enrollment.In total,566 consecutive patients from The First Affiliated Hospital of Zhejiang University were enrolled between January and March 2018.Urine samples were taken from all patients,82.7%(468/566)of which were hospitalized,and 17.3%(98/566)were outpatients.All samples were collected from adults aged 16-94-years-old(mean:60 years),and 220 specimens were from female patients and 346 were from males.Specimens were collected with the voided midstream technique in sterile containers and were processed within 2 hour(h).Each specimen was divided into two sterile tubes,one used for urine culture and the other for UF-5000 analysis.Pregnant woman,children,infants and catheterized patients were not enrolled to enhance the homogeneity in the participants.To avoid erratic results,aggregates or visible particle-containing specimens had also been excluded.Urine cultureQuantitative urine cultures were performed using well-mixed specimens and a 10-?L calibrated loop on Columbia blood agar plates.After 18-24 h acrobic incubation at 37?,the numbers of colonies were counted and multiplied by 103 to calculate one or two potentially pathogenic microorganisms per milliliter.In this study.cultures were considered significantly negative if bacterial growth was<104 CFU/mL.The VITKE2-Compact automated system(Bio-Merieux,Marcy-l'Etoile,Fraace)was used to identify bacteria.Samples showing growth of more than two colony species and without a predominant one were classified as culture positive but contaminated and were not subjected to the identification procedure.UF-5000 analysisAll urine specimens were analyzed for bacteria(BACT)and blood cell(WBC)using a UF-5000 analyzer(Sysmex Corporation,Kobe,Japan)after inoculating the culture.Each analysis was completed within 2 h of sample collection.ResultsAmong the 566 urine cultures,161 were positive(28.4%)and 405 were negative(71.6%).Among the 161 positive samples,30 samples positive for yeast were excluded and the remaining 131 were positive for bacteria.Among the 131 samples positive for bacteria,a single species was isolated in 87 samples,two species were isolated in 6 samples and 38 showed the growth of more than two species without a predominant one;the 38 samples with mixed flora were not subjected to the identification procedure.The bacteria identified(93 in total),in descending order of prevalence,were:Escherichia coli(n=33),Klebsiella pneumoniae(n=16),Proteus mirabilis(n=14),Enterococcus faecalis(n=8),Staphylococcus aureus(n=5),Enterococcus faecium(n=5),Pseudomonsa aeruginosa(n=5),Staphylococcus epidermidis(n=21,Streptococcus agalactiae(n=2),Streptococcus anginosus(n=2),Gardnerella vaginalis(n=2),Citrobacter freundii(n=1),Streptococcus mitis(n=1)and other bacteria(n=3);for yeast(30 in total),the order were:Candida albicans(n=14),Candida glabrata(n=8),Candida tropicalis(n=4),Candida guilliermondii(n=2)and other Candida species(n=2).Urine cultures were considered positive if the bacteria growth was?104 CFU/mL.Next,the UF-5000 BACT and WBC counts were compared with the urine culture results using ROC curve analysis.The area under the curve(AUC)for BACT was 0.961(95%confidence interval[CI]=0.940-0.976),which was greater than that of WBC(AUC=0.769,95%CI=0.729-0.805).Furthermore,using any cut-off value,the corresponding sensitivity and specificity were higher for BACT than WBC count.The performance of UF-5000 was examined using different cut-off values of the BACT count,then SEN,SPE,PPV and NPV,TP,FP and negative cases were calculated from the ROC curve analysis.In screening tests,SEN is the most important measure.According to the Guidance on the Application of Accreditation Criteria for Medical Laboratory Quality and Competence in the Field of Body Fluid Examination(CL41)issued by China National Accreditation Service for Confonnity Assessment(CNAS)in 2013,the FN rate should be<5%for automated urinalysis.Therefore,the optimal BACT count cut-off point was selected at 30/?L,with an optimal compromise SEN of 95.7%and SPE of 76.8%from the ROC analysis in all specimens.The highest SEN(97.8%)cut-off point for BACT count(15/?L)was excluded due to low SPE(64.2%).When considering the FN rates and NPV,the rest of the cut-off values were inappropriate for discriminating culture-negative urine specimens.Thus,the BACT cut-off value of 30/?L was the best,and had relatively high NPV(98.7%)and low numbers of FNs(4).Considering the gender factors,the AUC for male,was 0.976(95%CI=0.953-0.990),which was greater than that of female(AUC.931,95%CI=0.884-0.963).The male and female cut-off values of 30/?L was the best for male and for the female 28/?L,and their FN rate<5%.WBC count is an important factor for the clinical treatment of UTIs.Therefore,we combined BACT positivity(30/?L)with different WBC cut-off values and calculated SEN,SPE,PPV,NPV,TP and FP rates,and negative cases,which was supposed to maximize NPV and minimize FNs.When using this algorithm with the combined cut-off values of BACT<30/?L and/or WBC<200/?L,the UF-5000 demonstrated a SEN of 97.8%,a SPE of 74.6%,a PPV of 46.9%,a NPV of 99.3%,and an agreement of 78.9%with the culture method.Using this algorithm,only two FN specimens were obtained.The culture results for the two false negative samples were::Streptococcus mitis,(n=1,104 CFU/mL)and Streptococcus agalactiae(n=1,104 CFU/mL).Among the 93 cases with positive culture results,in terms of the performance of discriminating Gram-groups,the BACT-info was:Gram-positive in 19 cases,Gram-negative in 50 cases,Gram-pos/neg in 14 cases,and 10 cases were unclassified.In total,63(67.7%)cases were correctly discriminated by UF-5000,and it was clear that Gram-positive and Gram-negative BACT-info showed substantial agreement,with a Cohen's kappa concordance coefficient of 0.775(x2=31.65,p<0.001).Among the 14 Gram-pos/neg specimens,one was Gram-positive and 13 were Gram-neative;among the 10 unclassified specimens,six were Gram-positive and four were Gram-negative.The relationship between UF5000 and urine culture was acceptable for clinical use.The results of urine culture were divided into four grades:102103 CFU/mL,103-104 CFU/mL,104-105 CFU/mL and>105 CFU/mL,the median of UF5000 bacterial count was 2.0/?L(95%CI=1.6-23.0),7.8/?L(95%CI=2.2-40.7),47.1/?L(95%CI=39.5-303.3),8022.0/?L(95%CI=16110.9-25996.1),respectively There was a good correlation between UF5000 bacterial count results and standard urine culture results.ConclusionsThe new UF-5000 automated urine particle analyzer can provide reliable WBC and bacteria count results and information that can suggest bacterial Gram-groups,all of which are available in a few minutes and require no additional equipment or technologist.These preliminary results suggest that using the UF-5000 to rapidly discriminate culture-negative specimens will improve the quality of patient care by reducing the unnecessary initiation of antimicrobial treatment Additionally,the UF-5000 can significantly reduce the workload and costs of clinical laboratories.In conclusion,the performance of UF-5000 at rapid discriminating culture-negative specimens was quite acceptable for clinical use.