- Email: [email protected]
Does flow cytometry have a role in preliminary differentiation between urinary tract infections sustained by gram positive and gram negative bacteria? An Italian polycentric study
Clinica Chimica Acta 440 (2015) 152–156
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Does flow cytometry have a role in preliminary differentiation between urinary tract infections sustained by gram positive and gram negative bacteria? An Italian polycentric study Gianluca Gessoni a,⁎, Graziella Saccani b, Sara Valverde a, Fabio Manoni c, Marco Caputo d a
Clinical Pathology Dept., Madonna della Navicella Hospital, Chioggia, VE, Italy Clinical Chemistry and Microbiology Laboratory, Orlandi Hospital, Bussolengo, VR, Italy c Clinical Pathology Dept., Civil Hospital, Monselice, PD, Italy d Clinical Chemistry and Microbiology Laboratory, Fracastoro Hospital, S. Bonifacio, VR, Italy b
a r t i c l e
i n f o
Article history: Received 6 April 2014 Received in revised form 20 November 2014 Accepted 20 November 2014 Available online 26 November 2014 Keywords: Bacteriuria B_FSC UF-1000i Gram discrimination UTI
a b s t r a c t Background: Urine culture is the most frequently requested test for a Microbiology Lab. A reliable screening tool would be of paramount importance both to clinicians and laboratorians, provided that it could get fast and accurate negative results in order to rule-out urinary tract infection (UTI). Materials and methods: We evaluated 1907 consecutive urine samples from outpatients. Culture was performed on chromogenic agar with 1 μL loop, using 105 CFU/mL as a limit of positive growth. Using Sysmex Uf-1000i analyzer we evaluated bacteria forward scatter (B_FSC) and fluorescent light scatter (B_FLH) in a preliminary discrimination step for UTI caused by Gram+ or Gram− bacteria. Results: We got 512 positive samples. A mono-microbial infection was observed in 490 samples; two bacterial strains were isolated in 22 samples, so 534 bacterial strains were found: 392 Gram−, 133 Gram+ and 9 yeasts. Comparing Gram+ and Gram− bacteria we observed a statistically significant difference for B_FSC but not for B_FLH. In this application experimental cut-off value for B_FSC was 25ch. Using this cut-off to perform a presumptive identification of UTI sustained by Gram-+ bacteria, we observed a SE 0.68, SP 0.84. Conclusion: Our data although preliminary suggest that B_FSC could be useful in presumptive exclusion of UTI caused by Gram-positive bacteria. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Among transmissible diseases, urinary tract infections (UTI) are only second in frequency to upper respiratory tract infections, and urine culture is the most common bacteriological test in a clinical laboratory [1]. However, a number of unresolved problems still exist. The overall yield of positive culture results, even from patients with typical symptoms of UTI, is low despite the heavily labor- and timeconsuming procedures [2]. The generally accepted definition of significant bacteriuria in voided urine specimens is ≥105 CFU/mL of a single Abbreviations: AUC, Areas Under Curve; BACT, Bacteria; B_FSC, Bacteria Forward Scatter; B_FLH, Bacteria Fluorescent Light Scatter; CFU, Colony Forming Units; CI, 95% Confidence Interval; CPS ID3 agar, Chromogenic Agar Plates; DA, Diagnostic Accuracy; FN, False Positive; FP, False Negative; IQR, Inter Quartile Range; LEU, Leukocytes; ME, Median; NPV, Negative Predictive Value; PPV, Positive Predictive Value; ROC, Receiver Operating Characteristic; SE, Sensitivity; SP, Specificity; TN, True Negative; TP, True Positive; UTI, Urinary Tract Infections ⁎ Corresponding author at: Clinical Pathology, Madonna della Navicella Hospital, Strada Madonna Marina 500, 30015 Chioggia, VE, Italy. Tel.: +39 041 5534 400; fax: +39 041 5534 401. E-mail addresses: [email protected], [email protected] (G. Gessoni).
http://dx.doi.org/10.1016/j.cca.2014.11.022 0009-8981/© 2014 Elsevier B.V. All rights reserved.
microorganism [3], but lower limits were suggested for children, men, patients with underlying diseases, or when “fastidious” microorganisms are involved [4]. In the vast majority of patients, UTI are caused by Gram negative bacteria, (Enterobacteriaceae first, e.g. Escherichia coli, then nonfermenting Gram negative rods, such as Pseudomonas spp.); Gram positive bacteria (Enterococcus spp., Streptococcus spp. and Staphylococcus spp.) are involved in about 25% UTI [5,6]. In patients with UTI, the Turnaround Time (TAT) for a test result is no less than 48 h: you simply cannot get a 2-day time-span before starting antibacterial treatment in a symptomatic patient. So physicians usually start a blind, empiric therapy based on the sensitivities to chemotherapeutic agents known to be active against the bacteria most commonly involved in UTI. Unfortunately, the most active agents against Gram-negative bacteria are not very effective against Grampositive bacteria: an idea of the Gram characteristic of the germ involved in suspected UTI would certainly enhance the efficacy of empirical therapy [7–11]. Some evidence exists that the evaluation of “dimensional parameters” derived from the distribution histograms in the bacterial channel (bacteria forward scatter: B_FSC) of the modern cytometers can be useful in a rough, but extremely rapid etiological
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
differentiation [12]. The aim of this retrospective study was to evaluate the feasibility of a rapid presumptive identification of UTI caused by Gram negative bacteria using bacteria forward scatter (B_FSC) and bacteria fluorescent light scatter (B_FLH). 2. Materials and methods Three hospital-based clinical laboratories were involved in this study, according to the prerequisites of using dip-stick automated analyzers for routine chemical urinalysis and a flow cytometer Sysmex Uf-1000i for formed particle examination.
153
Variance analysis was expressed by the interquartile range (IQR). A Kruskal–Wallis test was performed for comparison of data. Receiver operating characteristic (ROC) curves were drawn by plotting sensitivity versus 1-specificity, to define the best cut-off values the Youden index was evaluated, and the areas under curve (AUCs) were measured. Finally, specificity (SP), sensitivity (SE), positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy (DA) were calculated.
3. Results
2.1. Sample collection
3.1. Microbiological analyses
We considered 1907 adult outpatients (775 males and 1132 females, age 18–70 years) submitted to our institutions for suspected UTI. Urines were collected in a sterile container (100 mL), fully equipped for sampling by vacuum tubes (Vacutest Kima, Arzergrande PD, Italy). Two separate tubes without preservative solution were immediately sampled, one for microbiological examination and one for Sysmex UF-1000i examination and kept refrigerated until analysis. Plate inoculation and Sysmex UF-1000i analysis were performed within 4 h from sample collection [13,14].
Bacterial isolates of this study are reported in Table 1. 553 out of 1907 samples (29%) showed a bacterial growth ≥ 105 CFU/mL. A mono-microbial infection was observed in 490 samples: 362 Gram negative, 119 Gram positive and 9 yeasts; two bacterial strains were isolated in 22 samples (8 had two Gram negative bacteria; 14 a Gram positive plus a Gram negative), 41 patients showed a poly-microbial flora and were considered as contaminated.
2.2. UF-1000i analysis Samples were processed on Sysmex UF-1000i Analyzer (Dasit, Milano, Italy). Briefly, a flow cytometer that counts, separates and analyzes microscopic particles suspended in a fluid stream. It also performs simultaneous, physico-chemical, multi-parametric analyses on single cells flowing through a detection system, in order to obtain adequate classification of urinary particles. The measured parameters are converted into electric signals, and the signal analysis enables classification and quantitation of each particle accordingly. All measurements are shown as a scattergram by means of a software (version 0018). Particle counts include erythrocytes, WBCs, epithelial cells, casts, bacteria, crystals and yeasts. UF-1000i has a separate analytical channel for bacteria, where urine specimen is mixed at 42 °C to a diluent that increases cell wall permeability and allows specific staining of bacterial nucleic acids with a dedicated polyethnic fluorescent dye. Particles are classified and quantified by considering their size- (impedance) and stainingcharacteristics using the forward scatter and the intensity of fluorescent light. Two additional parameters are available in this channel: the bacteria forward scatter (B_FSC) and the bacteria fluorescent light scatter (B_FLH), reported in arbitrary units (analytical channel — ch) and providing information about size (B_FSC) and nucleic acid contents (B_FLH).
3.2. Presumptive differentiation between Gram positive and Gram negative bacteria by UF-1000i For B_FSC and B_FLH median and quartiles have been calculated, as reported in Table 2, by considering Gram positive alone (119), Gram negative alone (370), a Gram positive and a Gram negative strain (mixed) (14), yeast (9), and contaminated samples (41). In Gram negative bacteria B_FSC median value was 20.60ch (1st quartile = 14.50ch, 3rd quartile = 33.03ch, IQR 18.53ch). In Gram positive bacteria B_FSC median value was 40.70ch (1st quartile = 27.80ch, 3rd quartile = 59.93ch, IQR 32.13ch). A high statistically significant difference (p b 0.001) in B_FSC between Gram positive and Gram negative was observed. A lower, but still significant difference (p b 0.01) in B_FSC remained between Gram negative bacteria and yeast and contaminated samples but not in samples with mixed growth of Gram positive and Gram negative bacteria (data reported in Table 2). Gram negative bacteria had a B_FLH median value = 85.50ch (1st quartile 77.62ch, 3rd quartile 112.35ch, IQR 34.73ch). B_FLH median value for Gram positive bacteria = 98.00ch (1st quartile 71.30ch, 3rd quartile 115.77ch, IQR 28.43ch). Here no statistically significant difference was detected (p N 0.5), Scattergram patterns from B_FSC channel, for Gram positive, Gram negative, association of a Gram positive plus a Gram negative and contaminated samples are reported in Fig. 1.
2.3. Microbiological analysis Quantitative urine culture was performed by using a 1 mL inoculation loop. Urine samples were routinely cultured for pathogens using the commercial chromogenic agar medium CPS ID3, (Biomerieux, Milano, Italy). Culture plates were incubated aerobically at 35 °C for 18–24 h. Quantification, in CFU/mL, was obtained multiplying the colonies numbered on the agar plate by the dilution factor. The culture was labeled as positive if containing ≥ 105 CFU/mL [15,16]. Standard biochemical identification and susceptibility tests to anti-microbic drugs were performed by using Vitek 2 analyzer (Biomerieux, Milano, Italy) [15]. 2.4. Statistical analysis Statistical tests were performed using a dedicated software (Analyse-it© version 2.03). A nonparametric statistical approach was adopted, evaluating median and 90% confidence intervals (CI 90%),
Table 1 Isolated strains from positive (≥105 CFU/mL) urine samples. Strains Candida spp. Citrobacter spp. Enterobacter spp. Enterococcus spp. Escherichia coli Klebsiella spp. Morganella spp. Proteus spp. Pseudomonas spp. Staphylococcus spp. Streptococcus spp.
9 11 11 63 279 53 4 11 23 29 41
1907 consecutive samples from adult outpatients. 553 samples. In 490 samples was observed a mono-microbic infection, in 22 samples were detected two pathogens and in 41 samples a polymicrobic growth was observed, these samples were considered as contaminated.
154
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
Table 2 Distribution of B_FSC and B-FLH in urinary tract infections sustained by Gram negative, Gram positive, mixed Gram positive and Gram negative bacteria and yeasts. B_FSC (ch)
Min
1st quartile Median 3rd quartile Max
Gram negative 6.80 Gram positive 18.20 Gram positive and Gram 9.80 negative Yeast 39.50 Mixed flora (contamination) 9.30
14.50 27.80 16.10
20.60 40.70 23.60
33.03 59.93 37.92
90.0 90.3 42.1
64.90 28.30
79.40 37.60
94.07 48.80
110.2 76.40
Gram negative Gram positive Gram positive and Gram negative Yeast Mixed flora (contamination)
77.62 87.13 90.82
85.50 98.00 101.50
112.35 115.57 112.80
214.10 139.80 120.60
91.90 101.63 71.40 92.70
119.40 105.90
133.40 115.70
148.60 156.80
70.10 71.30 84.40
B_FSC = bacteria forward scatter, and B_FLH = bacteria fluorescent light scatter. We observed in 370 samples Gram negative bacteria, in 119 Gram positive bacteria, in 14 Gram positive, plus Gram negative bacteria, in 9 yeasts (Candida spp.) and 41 contaminated samples (≥3 bacterial strains). Statistical significance for B_FSC: Gram negative bacteria versus Gram positive: p b 0.001; vs Yeasts: p b 0.01; vs contaminated samples: p b 0.01; vs associated Gram positive and plus Gram negative bacteria: p = 0.6. Statistical significance for B_FLH: Gram negative bacteria versus Gram positive: p b 0.01; vs Yeasts: p b 0.05; vs contaminated samples: p b 0.05; vs associated Gram positive and plus Gram negative bacteria: p = 0.7.
Fig. 2. ROC curve for B_FSC and B_FLH obtained by using the UF-1000i analyzer. ROC curves have been calculated by considering all samples with bacterial growth over 105 CFU/mL at quantitative culture (553). Gram positive alone (119), Gram negative alone (362), a Gram positive and a Gram negative strain (mixed) (22), yeast (9), samples considered as contaminated because of a polymicrobic growth (41).
ROC curve plots (Fig. 2) set for B_FSC an experimental cutoff value = 25ch, (Youden index 0.52, AUC = 0.78). The cut-off value for B-FLH was 90ch (Youden index 0.34, AUC = 0.64). This difference was statistically significant (p b 0.001).
Overall, 288 out of 553 positive samples (52%) showed B_FSC b 25ch; in 90% of these samples were observed Gram negative bacteria, in 3% Gram positive bacteria, in 2% a Gram positive plus a Gram negative bacteria, and in 5% a contamination.
Fig. 1. Scattergrams for B_FSC versus B-FLH and ch. Gram neg: Escherichia coli, Klebsiella spp., Pseudomonas spp.; Gram pos: Enterococcus spp., Staphylococcus spp., Streptococcus spp. B_FSC = bacteria forward scatter, B_FLH = bacteria fluorescent light scatter ch = arbitrary channel.
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
Instead 265 samples (48%) showed B_FSC N25ch; in 46% of these samples were observed Gram negative bacteria, in 31% Gram positive bacteria, in 2% a Gram positive plus a Gram negative bacteria, in 3% a yeast, and in 18% a contamination. Moreover 258/362 (71%) of UTI sustained by Gram negative bacteria showed a mean B_FSC b25ch and 108/119 (91%) of UTI sustained by Gram positive bacteria showed a mean B_FSC N25ch. As reported in Table 3 in presumptive rapid identification of UTI sustained by Gram negative bacteria B_FSC showed a SE of 0.68, a SP of 0.84, a PPV of 0.90, a NPV of 0.75, a DA of 0.69; B_FLH showed a SE of 0.57, a SP of 0.77, a PPV of 0.80, a NPV of 0.55, and a DA of 0.63. Distribution for B_FSC and B_FLH in Gram positive, Gram negative, yeast, Gram positive plus Gram positive UTI and contaminated samples is reported in Fig. 3. 4. Discussion A number of rapid methods for UTI diagnosis have been proposed in the past years: microscopic observation of untreated or stained samples; enzymatic methods using catalase, glucose oxidase, nitrate reductase, leukocyte esterase; filtration-based colorimetric methods, bioluminescence assays, photometric growth detection and, recently, flow cytometry [16–25]. In our study, disease prevalence (29%) and strain distribution of isolated microorganisms are consistent with data from the medical literature [4–11]. When deciding about an empirical treatment in UTI, Gram properties of the suspected germ would be valuable in selecting the most effective agent [7–11,26–28]. UF-1000i has a dedicated analytical flow channel named “BACT channel”, where specific reagents and algorithms provide data for bacteria detection and counting, but also a number of additional parameters are available, such as B_FSC, potentially useful in differentiating Gram-positive and Gram-negative bacteria [12,29]. The rationale for its use is that Gram negative bacteria do not usually aggregate to form clusters or chains, and tend to remain in suspension as single cells. Gram positive bacteria, instead, incline to arrange in chains (streptococci and enterococci) or clusters of different, greater size (staphylococci). The forward scatter measured in the bacteria channel of Sysmex UF-1000i analyzer effectively detects this difference [12,28], and the parameter correlates with the linear dimensions of the particles examined. As shown in Fig. 1, comparing Gram negative bacteria such as E. coli to Gram-positive such as Enterococcus spp., we got very different graphs. Using a two dimension scattergram reporting the B_FSC (y axis) and B_FLH (x axis), vertical distribution of E. coli is much less extended than Enterococcus spp. This is confirmed by the graph of the dimensional distribution showing a narrow curve for E. coli and a wider area for Enterococcus spp. In the same figure were reported examples of graphs observed in a mixed culture (E. coli plus Enterococcus spp.). It seems relevant to note that in two dimension scattergram it is possible to appreciate two different clusters
155
attributable to the two bacterial strains. The dimensional distribution scattergram shows a curve for B_FSC intermediate between those observed for Gram positive and Gram negative bacteria. Finally, scattergrams of samples considered as “contaminated” show an irregular and very broad distribution in both graphs. Camporese et al. [12] suggested to set a cut off value ≤ 30ch when using B_FSC in identifying Gram positive strains. In the present study, we found that a lower threshold, namely ≤ 25ch, gave an increase in SE and SP (0.68 and 0.84, respectively) and a better Youden index (0.52), compared to the ≤ 30ch (SE 0.73, SP 0.66, Youden index 0.395). In our population, a B_FSC b25ch could reliably rule out the great majority (91%) of UTI sustained by Gram positive bacteria. Moreover B_FSC with a cut-off at 25ch showed, in these patient's series, good SE (0.68), SP (0.84) and PPV (0.90) in presumptive identification of UTI sustained by Gram negative bacteria. This study presented some limits. For example the retrospective design: data relating to B-FLH and B_FSC were extracted and processed from a data base only for samples with bacterial growth ≥ 105 CFU/mL. We could not test samples with “false positive” results generated by Sysmex UF-1000i. This should be kept in mind in transferring our experiment in diagnostic routine because contaminated and polymicrobial UTI cannot be identified at the moment of analysis so ROC-curve data of this study cannot be transferred to this situation. Other limits were the patient's selection (only outpatients were considered) and the lack of information on the possible antibacterial therapy administered patients before microbiological examination. These may have been interesting because one could expect that with antibiotic therapy, the shape of the bacteria could be influenced and this might influence the results or the typical scattergrams. On the other hand there are strong points in proposing a Sysmex UF-1000i analyzer for rapid diagnosis of UTI. For example this analyzer is capable of providing reliable quantitative results on bacteriuria and leukocyturia, moreover these data will be available in a few minutes and in complete automation by a standard urinalysis, with no need for additional equipment, any additional waste of human and technological resources. In a laboratory where for the collection of urine samples are adopted sterile, standardized containers, fully equipped for sampling by vacuum tubes, you can make a first tube for screening with the Sysmex UF-1000i analyzer, making the second tube, for sending in Microbiology area only for screening positive samples saves time and resources and reducing the number of samples to be run. If confirmed in studies with larger series, these preliminary data can suggest the use of flow cytometers in the rapid diagnosis of UTI as a means for adding clinical value to the laboratory report by rapid presumptive identification of infections sustained by Gram negative bacteria. In conclusion, we support the rationale of flow cytometry as a fast and reliable screening tool in the diagnosis of UTI [12,24]. In our opinion, additional studies in due course will help in deciding about
Table 3 Sensitivity, specificity, positive and negative predictive values and diagnostic accuracy for discrimination between urinary tract infections sustained by Gram positive and Gram negative bacteria.
B_FSC B_FLH
SE
SP
PPV
NPV
DA
0.68 (0.61–0.75) 0.57 (0.52–0.63)
0.84 (0.75–0.91) 0.77 (0.67–0.87)
0.90 (0.86–0.94) 0.84 (0.79–0.89)
0.75 (0.71–0.79) 0.45 (0.41–0.49)
0.69 (0.65–0.74) 0.63 (0.59–0.67)
(95% Confidence Intervals). B_FSC = bacteria forward scatter, B_FLH = bacteria fluorescent light scatter. B_FSC: cut-off value = 25ch; AUC = 0.78. B_FLC: cut-off value = 90ch; AUC = 0.64. Se = Sensitivity, SP = Specificity, PPV = Positive Predictive Value, NPV = Negative Predictive Value, DA = Diagnostic Accuracy.
156
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
Fig. 3. B_FSC and B_FLH in urinary tract infections. In this figure reported distribution of B_FSC and B_FLH in urinary tract infections were sustained by Gram negative bacteria (370), Gram positive bacteria (119), an association of Gram positive and Gram negative bacteria (14), and Yeast (9). Moreover reported samples were considered as contaminated because of a polymicrobic growth (41).
the routine use of this technology as a further step forward in terms of evidence-based cost savings, antibiotic stewardship and, finally, better overall healthcare outcomes. Conflict of interest The authors of this manuscript declare that they have no conflicts of interest (financial and/or commercial). References [1] Schifman R, Wieden M, Brooker J, et al. Bacteriuria screening by direct bioluminescence assay of ATP. J Clin Microbiol 1984;20:644–8. [2] Wu T, Williams E, Koo S, McLowry J. Evaluation of three bacteriuria screening methods in a clinical research hospital. J Clin Microbiol 1985;21:796–9. [3] Kass E. Asymptomatic infections of the urinary tract. Trans Assoc Am Phys 1956;69: 56–63. [4] Pappas G. Laboratory in the diagnosis and management of urinary tract infections. Med Clin North Am 1991;75:313–25. [5] Genao L, Buhr GT. Urinary tract infections in older adults residing in long-term care facilities. Ann Longterm Care 2012;20:33–8. [6] Jadhav SV, Sardar M, Vyawahare C, Misra RR. The spectrum and aetiology of mycotic infections from a tertiary care hospital from Western part of India. J Clin Diagn Res 2013;7:2157–9. [7] Morrissey I, Hackel M, Badal R, Bouchillon S, Hawser S, Biedenbach D. A review of ten years of the Study for Monitoring Antimicrobial Resistance Trends (SMART) from 2002 to 2011. Pharmaceuticals (Basel) 2013;6:1335–46. [8] Schimemann G, Gágyor I, Hummers-Pradier E, Bleidorn J. Resistance profiles of urinary tract infections in general practice, an observational study. BMC Urol 2012;12: 33–7. [9] McQuiston Haslund J, Rosborg Dinesen M, Sternhagen Nielsen AB, Llor C, Bjerrum L. Different recommendations for empiric first-choice antibiotic treatment of uncomplicated urinary tract infections in Europe. Scand J Prim Health Care 2013;31: 235–40. [10] Lo DS, Shieh HH, Ragazzi SL, Koch VH, Martinez MB, Gilio AE. Community-acquired urinary tract infection: age and gender-dependent etiology. J Bras Nefrol 2013;35: 93–8. [11] Cunha BA, Schoch PE, Hage JR. Nitrofurantoin: preferred empiric therapy for community-acquired lower urinary tract infections. Mayo Clin Proc 2011;86: 1243–4. [12] De Rosa R, Grosso S, Bruschetta G, et al. A. Evaluation of the Sysmex UF1000i flow cytometer for ruling out bacterial urinary tract infection. Clin Chim Acta 2010;411: 1137–42. [13] Manoni F, Gessoni G, Alessio MG, et al. Mid-stream vs first-voided urine collection by using automated analyzers for particle examination in healthy subjects: an Italian multi center study. Clin Chem Lab Med 2001;50:679–84.
[14] Manoni F, Caleffi A, Gessoni G, et al. L'esame chimico, morfologico e colturale delle urine: proposta di linee guida per una procedura standardizzata della fase pre analitica. Biochim Clin 2011;5:131–9. [15] Munoz-Dávila MJ, Roig M, Yagüe G, Blázquez A, Salvador C, Segovia M. Comparative evaluation of Vitek 2 identification and susceptibility testing of urinary tract pathogens directly and isolated from chromogenic media. Eur J Clin Microbiol Infect Dis 2013;32:773–80. [16] Bartlet R, Galen R. Predictive value of urine culture. J Infect Dis 1983;79:179–82. [17] Pezlo M, Amsterdam D, Anhalt J, et al. Detection of bacteriuria and pyuria by URISCREEN a rapid enzymatic screening test. J Clin Microbiol 1992;30:680–4. [18] Pezlo M. Detection of urinary tract infections by rapid methods. Clin Microbiol Rev 1988;1:268–80. [19] Koenig C, Tick L, Hanna B. Analyses of the flash track DNA probe and UTIscreen bioluminescence tests for bacteriuria. J Clin Microbiol 1992;30:342–5. [20] Marschal M, Wienke M, Hoering S, Autenrieth IB, Frick JS. Evaluation of 3 different rapid automated systems for diagnosis of urinary tract infections. Diagn Microbiol Infect Dis 2012;72:125–30. [21] Milagro A, Moles B, Seoane A, Navascués J, Blasco M, García-Moya JB. UTIscreen versus UROQUICK: two semiautomatic systems for bacteriuria detection. Enferm Infecc Microbiol Clin 1999;17:398–400. [22] Murray P, Smith T, McKinney R. Clinical evaluation of three urine screening tests. J Clin Microbiol 1987;25:467–70. [23] Koken T, Aktepe O, Sertesr M, Samli M, Kahraman A, Dogan N. Determination of cutoff values for leukocytes and bacteria for urine flow cytometer (UF-100) in urinary tract infection. Int Urol Nephrol 2002;34:175–8. [24] Gutiérrez-Fernández J, Lara A, Bautista MF, et al. Performance of the Sysmex UF1000i system in screening for significant bacteriuria before quantitative culture of aerobic/ facultative fast-growth bacteria in a reference hospital. J Appl Microbiol 2012;113: 609–14. [25] Falbo R, Sala R, Signorelli S, Venturini N, Sisnorini S, Brambilla P. Bacteriuria screening by automated whole-field-image-based microscopy reduces the number of necessary urine cultures. J Clin Microbiol 2012;50:1427–9. [26] Hayami H, Takahashi S, Ishikawa K, et al. Nationwide surveillance of bacterial pathogens from patients with acute uncomplicated cystitis conducted by the Japanese surveillance committee during 2009 and 2010: antimicrobial susceptibility of Escherichia coli and Staphylococcus saprophyticus. J Infect Chemother 2013;19: 393–403. [27] Neuzillet Y, Naber KG, Schito G, Gualco L, Botto H. French results of the ARESC study: clinical aspects and epidemiology of antimicrobial resistance in female patients with cystitis. Implications for empiric therapy. Med Mal Infect 2012;42:66–75. [28] Parish A, Holliday K. Long-term care acquired urinary tract infections' antibiotic resistance patterns and empiric therapy: a pilot study. Geriatr Nurs 2012;33:473–8. [29] Wada A, Kono M, Kawauchi S, Takagi Y, Morikawa T, Funakoshi K. Rapid discrimination of Gram-positive and Gram-negative bacteria in liquid samples by using NaOHsodium dodecyl sulfate solution and flow cytometry. PLoS ONE 2012;7:e47093. http://dx.doi.org/10.1371/journal.pone.0047093.
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Does flow cytometry have a role in preliminary differentiation between urinary tract infections sustained by gram positive and gram negative bacteria? An Italian polycentric study Gianluca Gessoni a,⁎, Graziella Saccani b, Sara Valverde a, Fabio Manoni c, Marco Caputo d a
Clinical Pathology Dept., Madonna della Navicella Hospital, Chioggia, VE, Italy Clinical Chemistry and Microbiology Laboratory, Orlandi Hospital, Bussolengo, VR, Italy c Clinical Pathology Dept., Civil Hospital, Monselice, PD, Italy d Clinical Chemistry and Microbiology Laboratory, Fracastoro Hospital, S. Bonifacio, VR, Italy b
a r t i c l e
i n f o
Article history: Received 6 April 2014 Received in revised form 20 November 2014 Accepted 20 November 2014 Available online 26 November 2014 Keywords: Bacteriuria B_FSC UF-1000i Gram discrimination UTI
a b s t r a c t Background: Urine culture is the most frequently requested test for a Microbiology Lab. A reliable screening tool would be of paramount importance both to clinicians and laboratorians, provided that it could get fast and accurate negative results in order to rule-out urinary tract infection (UTI). Materials and methods: We evaluated 1907 consecutive urine samples from outpatients. Culture was performed on chromogenic agar with 1 μL loop, using 105 CFU/mL as a limit of positive growth. Using Sysmex Uf-1000i analyzer we evaluated bacteria forward scatter (B_FSC) and fluorescent light scatter (B_FLH) in a preliminary discrimination step for UTI caused by Gram+ or Gram− bacteria. Results: We got 512 positive samples. A mono-microbial infection was observed in 490 samples; two bacterial strains were isolated in 22 samples, so 534 bacterial strains were found: 392 Gram−, 133 Gram+ and 9 yeasts. Comparing Gram+ and Gram− bacteria we observed a statistically significant difference for B_FSC but not for B_FLH. In this application experimental cut-off value for B_FSC was 25ch. Using this cut-off to perform a presumptive identification of UTI sustained by Gram-+ bacteria, we observed a SE 0.68, SP 0.84. Conclusion: Our data although preliminary suggest that B_FSC could be useful in presumptive exclusion of UTI caused by Gram-positive bacteria. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Among transmissible diseases, urinary tract infections (UTI) are only second in frequency to upper respiratory tract infections, and urine culture is the most common bacteriological test in a clinical laboratory [1]. However, a number of unresolved problems still exist. The overall yield of positive culture results, even from patients with typical symptoms of UTI, is low despite the heavily labor- and timeconsuming procedures [2]. The generally accepted definition of significant bacteriuria in voided urine specimens is ≥105 CFU/mL of a single Abbreviations: AUC, Areas Under Curve; BACT, Bacteria; B_FSC, Bacteria Forward Scatter; B_FLH, Bacteria Fluorescent Light Scatter; CFU, Colony Forming Units; CI, 95% Confidence Interval; CPS ID3 agar, Chromogenic Agar Plates; DA, Diagnostic Accuracy; FN, False Positive; FP, False Negative; IQR, Inter Quartile Range; LEU, Leukocytes; ME, Median; NPV, Negative Predictive Value; PPV, Positive Predictive Value; ROC, Receiver Operating Characteristic; SE, Sensitivity; SP, Specificity; TN, True Negative; TP, True Positive; UTI, Urinary Tract Infections ⁎ Corresponding author at: Clinical Pathology, Madonna della Navicella Hospital, Strada Madonna Marina 500, 30015 Chioggia, VE, Italy. Tel.: +39 041 5534 400; fax: +39 041 5534 401. E-mail addresses: [email protected], [email protected] (G. Gessoni).
http://dx.doi.org/10.1016/j.cca.2014.11.022 0009-8981/© 2014 Elsevier B.V. All rights reserved.
microorganism [3], but lower limits were suggested for children, men, patients with underlying diseases, or when “fastidious” microorganisms are involved [4]. In the vast majority of patients, UTI are caused by Gram negative bacteria, (Enterobacteriaceae first, e.g. Escherichia coli, then nonfermenting Gram negative rods, such as Pseudomonas spp.); Gram positive bacteria (Enterococcus spp., Streptococcus spp. and Staphylococcus spp.) are involved in about 25% UTI [5,6]. In patients with UTI, the Turnaround Time (TAT) for a test result is no less than 48 h: you simply cannot get a 2-day time-span before starting antibacterial treatment in a symptomatic patient. So physicians usually start a blind, empiric therapy based on the sensitivities to chemotherapeutic agents known to be active against the bacteria most commonly involved in UTI. Unfortunately, the most active agents against Gram-negative bacteria are not very effective against Grampositive bacteria: an idea of the Gram characteristic of the germ involved in suspected UTI would certainly enhance the efficacy of empirical therapy [7–11]. Some evidence exists that the evaluation of “dimensional parameters” derived from the distribution histograms in the bacterial channel (bacteria forward scatter: B_FSC) of the modern cytometers can be useful in a rough, but extremely rapid etiological
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
differentiation [12]. The aim of this retrospective study was to evaluate the feasibility of a rapid presumptive identification of UTI caused by Gram negative bacteria using bacteria forward scatter (B_FSC) and bacteria fluorescent light scatter (B_FLH). 2. Materials and methods Three hospital-based clinical laboratories were involved in this study, according to the prerequisites of using dip-stick automated analyzers for routine chemical urinalysis and a flow cytometer Sysmex Uf-1000i for formed particle examination.
153
Variance analysis was expressed by the interquartile range (IQR). A Kruskal–Wallis test was performed for comparison of data. Receiver operating characteristic (ROC) curves were drawn by plotting sensitivity versus 1-specificity, to define the best cut-off values the Youden index was evaluated, and the areas under curve (AUCs) were measured. Finally, specificity (SP), sensitivity (SE), positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy (DA) were calculated.
3. Results
2.1. Sample collection
3.1. Microbiological analyses
We considered 1907 adult outpatients (775 males and 1132 females, age 18–70 years) submitted to our institutions for suspected UTI. Urines were collected in a sterile container (100 mL), fully equipped for sampling by vacuum tubes (Vacutest Kima, Arzergrande PD, Italy). Two separate tubes without preservative solution were immediately sampled, one for microbiological examination and one for Sysmex UF-1000i examination and kept refrigerated until analysis. Plate inoculation and Sysmex UF-1000i analysis were performed within 4 h from sample collection [13,14].
Bacterial isolates of this study are reported in Table 1. 553 out of 1907 samples (29%) showed a bacterial growth ≥ 105 CFU/mL. A mono-microbial infection was observed in 490 samples: 362 Gram negative, 119 Gram positive and 9 yeasts; two bacterial strains were isolated in 22 samples (8 had two Gram negative bacteria; 14 a Gram positive plus a Gram negative), 41 patients showed a poly-microbial flora and were considered as contaminated.
2.2. UF-1000i analysis Samples were processed on Sysmex UF-1000i Analyzer (Dasit, Milano, Italy). Briefly, a flow cytometer that counts, separates and analyzes microscopic particles suspended in a fluid stream. It also performs simultaneous, physico-chemical, multi-parametric analyses on single cells flowing through a detection system, in order to obtain adequate classification of urinary particles. The measured parameters are converted into electric signals, and the signal analysis enables classification and quantitation of each particle accordingly. All measurements are shown as a scattergram by means of a software (version 0018). Particle counts include erythrocytes, WBCs, epithelial cells, casts, bacteria, crystals and yeasts. UF-1000i has a separate analytical channel for bacteria, where urine specimen is mixed at 42 °C to a diluent that increases cell wall permeability and allows specific staining of bacterial nucleic acids with a dedicated polyethnic fluorescent dye. Particles are classified and quantified by considering their size- (impedance) and stainingcharacteristics using the forward scatter and the intensity of fluorescent light. Two additional parameters are available in this channel: the bacteria forward scatter (B_FSC) and the bacteria fluorescent light scatter (B_FLH), reported in arbitrary units (analytical channel — ch) and providing information about size (B_FSC) and nucleic acid contents (B_FLH).
3.2. Presumptive differentiation between Gram positive and Gram negative bacteria by UF-1000i For B_FSC and B_FLH median and quartiles have been calculated, as reported in Table 2, by considering Gram positive alone (119), Gram negative alone (370), a Gram positive and a Gram negative strain (mixed) (14), yeast (9), and contaminated samples (41). In Gram negative bacteria B_FSC median value was 20.60ch (1st quartile = 14.50ch, 3rd quartile = 33.03ch, IQR 18.53ch). In Gram positive bacteria B_FSC median value was 40.70ch (1st quartile = 27.80ch, 3rd quartile = 59.93ch, IQR 32.13ch). A high statistically significant difference (p b 0.001) in B_FSC between Gram positive and Gram negative was observed. A lower, but still significant difference (p b 0.01) in B_FSC remained between Gram negative bacteria and yeast and contaminated samples but not in samples with mixed growth of Gram positive and Gram negative bacteria (data reported in Table 2). Gram negative bacteria had a B_FLH median value = 85.50ch (1st quartile 77.62ch, 3rd quartile 112.35ch, IQR 34.73ch). B_FLH median value for Gram positive bacteria = 98.00ch (1st quartile 71.30ch, 3rd quartile 115.77ch, IQR 28.43ch). Here no statistically significant difference was detected (p N 0.5), Scattergram patterns from B_FSC channel, for Gram positive, Gram negative, association of a Gram positive plus a Gram negative and contaminated samples are reported in Fig. 1.
2.3. Microbiological analysis Quantitative urine culture was performed by using a 1 mL inoculation loop. Urine samples were routinely cultured for pathogens using the commercial chromogenic agar medium CPS ID3, (Biomerieux, Milano, Italy). Culture plates were incubated aerobically at 35 °C for 18–24 h. Quantification, in CFU/mL, was obtained multiplying the colonies numbered on the agar plate by the dilution factor. The culture was labeled as positive if containing ≥ 105 CFU/mL [15,16]. Standard biochemical identification and susceptibility tests to anti-microbic drugs were performed by using Vitek 2 analyzer (Biomerieux, Milano, Italy) [15]. 2.4. Statistical analysis Statistical tests were performed using a dedicated software (Analyse-it© version 2.03). A nonparametric statistical approach was adopted, evaluating median and 90% confidence intervals (CI 90%),
Table 1 Isolated strains from positive (≥105 CFU/mL) urine samples. Strains Candida spp. Citrobacter spp. Enterobacter spp. Enterococcus spp. Escherichia coli Klebsiella spp. Morganella spp. Proteus spp. Pseudomonas spp. Staphylococcus spp. Streptococcus spp.
9 11 11 63 279 53 4 11 23 29 41
1907 consecutive samples from adult outpatients. 553 samples. In 490 samples was observed a mono-microbic infection, in 22 samples were detected two pathogens and in 41 samples a polymicrobic growth was observed, these samples were considered as contaminated.
154
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
Table 2 Distribution of B_FSC and B-FLH in urinary tract infections sustained by Gram negative, Gram positive, mixed Gram positive and Gram negative bacteria and yeasts. B_FSC (ch)
Min
1st quartile Median 3rd quartile Max
Gram negative 6.80 Gram positive 18.20 Gram positive and Gram 9.80 negative Yeast 39.50 Mixed flora (contamination) 9.30
14.50 27.80 16.10
20.60 40.70 23.60
33.03 59.93 37.92
90.0 90.3 42.1
64.90 28.30
79.40 37.60
94.07 48.80
110.2 76.40
Gram negative Gram positive Gram positive and Gram negative Yeast Mixed flora (contamination)
77.62 87.13 90.82
85.50 98.00 101.50
112.35 115.57 112.80
214.10 139.80 120.60
91.90 101.63 71.40 92.70
119.40 105.90
133.40 115.70
148.60 156.80
70.10 71.30 84.40
B_FSC = bacteria forward scatter, and B_FLH = bacteria fluorescent light scatter. We observed in 370 samples Gram negative bacteria, in 119 Gram positive bacteria, in 14 Gram positive, plus Gram negative bacteria, in 9 yeasts (Candida spp.) and 41 contaminated samples (≥3 bacterial strains). Statistical significance for B_FSC: Gram negative bacteria versus Gram positive: p b 0.001; vs Yeasts: p b 0.01; vs contaminated samples: p b 0.01; vs associated Gram positive and plus Gram negative bacteria: p = 0.6. Statistical significance for B_FLH: Gram negative bacteria versus Gram positive: p b 0.01; vs Yeasts: p b 0.05; vs contaminated samples: p b 0.05; vs associated Gram positive and plus Gram negative bacteria: p = 0.7.
Fig. 2. ROC curve for B_FSC and B_FLH obtained by using the UF-1000i analyzer. ROC curves have been calculated by considering all samples with bacterial growth over 105 CFU/mL at quantitative culture (553). Gram positive alone (119), Gram negative alone (362), a Gram positive and a Gram negative strain (mixed) (22), yeast (9), samples considered as contaminated because of a polymicrobic growth (41).
ROC curve plots (Fig. 2) set for B_FSC an experimental cutoff value = 25ch, (Youden index 0.52, AUC = 0.78). The cut-off value for B-FLH was 90ch (Youden index 0.34, AUC = 0.64). This difference was statistically significant (p b 0.001).
Overall, 288 out of 553 positive samples (52%) showed B_FSC b 25ch; in 90% of these samples were observed Gram negative bacteria, in 3% Gram positive bacteria, in 2% a Gram positive plus a Gram negative bacteria, and in 5% a contamination.
Fig. 1. Scattergrams for B_FSC versus B-FLH and ch. Gram neg: Escherichia coli, Klebsiella spp., Pseudomonas spp.; Gram pos: Enterococcus spp., Staphylococcus spp., Streptococcus spp. B_FSC = bacteria forward scatter, B_FLH = bacteria fluorescent light scatter ch = arbitrary channel.
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
Instead 265 samples (48%) showed B_FSC N25ch; in 46% of these samples were observed Gram negative bacteria, in 31% Gram positive bacteria, in 2% a Gram positive plus a Gram negative bacteria, in 3% a yeast, and in 18% a contamination. Moreover 258/362 (71%) of UTI sustained by Gram negative bacteria showed a mean B_FSC b25ch and 108/119 (91%) of UTI sustained by Gram positive bacteria showed a mean B_FSC N25ch. As reported in Table 3 in presumptive rapid identification of UTI sustained by Gram negative bacteria B_FSC showed a SE of 0.68, a SP of 0.84, a PPV of 0.90, a NPV of 0.75, a DA of 0.69; B_FLH showed a SE of 0.57, a SP of 0.77, a PPV of 0.80, a NPV of 0.55, and a DA of 0.63. Distribution for B_FSC and B_FLH in Gram positive, Gram negative, yeast, Gram positive plus Gram positive UTI and contaminated samples is reported in Fig. 3. 4. Discussion A number of rapid methods for UTI diagnosis have been proposed in the past years: microscopic observation of untreated or stained samples; enzymatic methods using catalase, glucose oxidase, nitrate reductase, leukocyte esterase; filtration-based colorimetric methods, bioluminescence assays, photometric growth detection and, recently, flow cytometry [16–25]. In our study, disease prevalence (29%) and strain distribution of isolated microorganisms are consistent with data from the medical literature [4–11]. When deciding about an empirical treatment in UTI, Gram properties of the suspected germ would be valuable in selecting the most effective agent [7–11,26–28]. UF-1000i has a dedicated analytical flow channel named “BACT channel”, where specific reagents and algorithms provide data for bacteria detection and counting, but also a number of additional parameters are available, such as B_FSC, potentially useful in differentiating Gram-positive and Gram-negative bacteria [12,29]. The rationale for its use is that Gram negative bacteria do not usually aggregate to form clusters or chains, and tend to remain in suspension as single cells. Gram positive bacteria, instead, incline to arrange in chains (streptococci and enterococci) or clusters of different, greater size (staphylococci). The forward scatter measured in the bacteria channel of Sysmex UF-1000i analyzer effectively detects this difference [12,28], and the parameter correlates with the linear dimensions of the particles examined. As shown in Fig. 1, comparing Gram negative bacteria such as E. coli to Gram-positive such as Enterococcus spp., we got very different graphs. Using a two dimension scattergram reporting the B_FSC (y axis) and B_FLH (x axis), vertical distribution of E. coli is much less extended than Enterococcus spp. This is confirmed by the graph of the dimensional distribution showing a narrow curve for E. coli and a wider area for Enterococcus spp. In the same figure were reported examples of graphs observed in a mixed culture (E. coli plus Enterococcus spp.). It seems relevant to note that in two dimension scattergram it is possible to appreciate two different clusters
155
attributable to the two bacterial strains. The dimensional distribution scattergram shows a curve for B_FSC intermediate between those observed for Gram positive and Gram negative bacteria. Finally, scattergrams of samples considered as “contaminated” show an irregular and very broad distribution in both graphs. Camporese et al. [12] suggested to set a cut off value ≤ 30ch when using B_FSC in identifying Gram positive strains. In the present study, we found that a lower threshold, namely ≤ 25ch, gave an increase in SE and SP (0.68 and 0.84, respectively) and a better Youden index (0.52), compared to the ≤ 30ch (SE 0.73, SP 0.66, Youden index 0.395). In our population, a B_FSC b25ch could reliably rule out the great majority (91%) of UTI sustained by Gram positive bacteria. Moreover B_FSC with a cut-off at 25ch showed, in these patient's series, good SE (0.68), SP (0.84) and PPV (0.90) in presumptive identification of UTI sustained by Gram negative bacteria. This study presented some limits. For example the retrospective design: data relating to B-FLH and B_FSC were extracted and processed from a data base only for samples with bacterial growth ≥ 105 CFU/mL. We could not test samples with “false positive” results generated by Sysmex UF-1000i. This should be kept in mind in transferring our experiment in diagnostic routine because contaminated and polymicrobial UTI cannot be identified at the moment of analysis so ROC-curve data of this study cannot be transferred to this situation. Other limits were the patient's selection (only outpatients were considered) and the lack of information on the possible antibacterial therapy administered patients before microbiological examination. These may have been interesting because one could expect that with antibiotic therapy, the shape of the bacteria could be influenced and this might influence the results or the typical scattergrams. On the other hand there are strong points in proposing a Sysmex UF-1000i analyzer for rapid diagnosis of UTI. For example this analyzer is capable of providing reliable quantitative results on bacteriuria and leukocyturia, moreover these data will be available in a few minutes and in complete automation by a standard urinalysis, with no need for additional equipment, any additional waste of human and technological resources. In a laboratory where for the collection of urine samples are adopted sterile, standardized containers, fully equipped for sampling by vacuum tubes, you can make a first tube for screening with the Sysmex UF-1000i analyzer, making the second tube, for sending in Microbiology area only for screening positive samples saves time and resources and reducing the number of samples to be run. If confirmed in studies with larger series, these preliminary data can suggest the use of flow cytometers in the rapid diagnosis of UTI as a means for adding clinical value to the laboratory report by rapid presumptive identification of infections sustained by Gram negative bacteria. In conclusion, we support the rationale of flow cytometry as a fast and reliable screening tool in the diagnosis of UTI [12,24]. In our opinion, additional studies in due course will help in deciding about
Table 3 Sensitivity, specificity, positive and negative predictive values and diagnostic accuracy for discrimination between urinary tract infections sustained by Gram positive and Gram negative bacteria.
B_FSC B_FLH
SE
SP
PPV
NPV
DA
0.68 (0.61–0.75) 0.57 (0.52–0.63)
0.84 (0.75–0.91) 0.77 (0.67–0.87)
0.90 (0.86–0.94) 0.84 (0.79–0.89)
0.75 (0.71–0.79) 0.45 (0.41–0.49)
0.69 (0.65–0.74) 0.63 (0.59–0.67)
(95% Confidence Intervals). B_FSC = bacteria forward scatter, B_FLH = bacteria fluorescent light scatter. B_FSC: cut-off value = 25ch; AUC = 0.78. B_FLC: cut-off value = 90ch; AUC = 0.64. Se = Sensitivity, SP = Specificity, PPV = Positive Predictive Value, NPV = Negative Predictive Value, DA = Diagnostic Accuracy.
156
G. Gessoni et al. / Clinica Chimica Acta 440 (2015) 152–156
Fig. 3. B_FSC and B_FLH in urinary tract infections. In this figure reported distribution of B_FSC and B_FLH in urinary tract infections were sustained by Gram negative bacteria (370), Gram positive bacteria (119), an association of Gram positive and Gram negative bacteria (14), and Yeast (9). Moreover reported samples were considered as contaminated because of a polymicrobic growth (41).
the routine use of this technology as a further step forward in terms of evidence-based cost savings, antibiotic stewardship and, finally, better overall healthcare outcomes. Conflict of interest The authors of this manuscript declare that they have no conflicts of interest (financial and/or commercial). References [1] Schifman R, Wieden M, Brooker J, et al. Bacteriuria screening by direct bioluminescence assay of ATP. J Clin Microbiol 1984;20:644–8. [2] Wu T, Williams E, Koo S, McLowry J. Evaluation of three bacteriuria screening methods in a clinical research hospital. J Clin Microbiol 1985;21:796–9. [3] Kass E. Asymptomatic infections of the urinary tract. Trans Assoc Am Phys 1956;69: 56–63. [4] Pappas G. Laboratory in the diagnosis and management of urinary tract infections. Med Clin North Am 1991;75:313–25. [5] Genao L, Buhr GT. Urinary tract infections in older adults residing in long-term care facilities. Ann Longterm Care 2012;20:33–8. [6] Jadhav SV, Sardar M, Vyawahare C, Misra RR. The spectrum and aetiology of mycotic infections from a tertiary care hospital from Western part of India. J Clin Diagn Res 2013;7:2157–9. [7] Morrissey I, Hackel M, Badal R, Bouchillon S, Hawser S, Biedenbach D. A review of ten years of the Study for Monitoring Antimicrobial Resistance Trends (SMART) from 2002 to 2011. Pharmaceuticals (Basel) 2013;6:1335–46. [8] Schimemann G, Gágyor I, Hummers-Pradier E, Bleidorn J. Resistance profiles of urinary tract infections in general practice, an observational study. BMC Urol 2012;12: 33–7. [9] McQuiston Haslund J, Rosborg Dinesen M, Sternhagen Nielsen AB, Llor C, Bjerrum L. Different recommendations for empiric first-choice antibiotic treatment of uncomplicated urinary tract infections in Europe. Scand J Prim Health Care 2013;31: 235–40. [10] Lo DS, Shieh HH, Ragazzi SL, Koch VH, Martinez MB, Gilio AE. Community-acquired urinary tract infection: age and gender-dependent etiology. J Bras Nefrol 2013;35: 93–8. [11] Cunha BA, Schoch PE, Hage JR. Nitrofurantoin: preferred empiric therapy for community-acquired lower urinary tract infections. Mayo Clin Proc 2011;86: 1243–4. [12] De Rosa R, Grosso S, Bruschetta G, et al. A. Evaluation of the Sysmex UF1000i flow cytometer for ruling out bacterial urinary tract infection. Clin Chim Acta 2010;411: 1137–42. [13] Manoni F, Gessoni G, Alessio MG, et al. Mid-stream vs first-voided urine collection by using automated analyzers for particle examination in healthy subjects: an Italian multi center study. Clin Chem Lab Med 2001;50:679–84.
[14] Manoni F, Caleffi A, Gessoni G, et al. L'esame chimico, morfologico e colturale delle urine: proposta di linee guida per una procedura standardizzata della fase pre analitica. Biochim Clin 2011;5:131–9. [15] Munoz-Dávila MJ, Roig M, Yagüe G, Blázquez A, Salvador C, Segovia M. Comparative evaluation of Vitek 2 identification and susceptibility testing of urinary tract pathogens directly and isolated from chromogenic media. Eur J Clin Microbiol Infect Dis 2013;32:773–80. [16] Bartlet R, Galen R. Predictive value of urine culture. J Infect Dis 1983;79:179–82. [17] Pezlo M, Amsterdam D, Anhalt J, et al. Detection of bacteriuria and pyuria by URISCREEN a rapid enzymatic screening test. J Clin Microbiol 1992;30:680–4. [18] Pezlo M. Detection of urinary tract infections by rapid methods. Clin Microbiol Rev 1988;1:268–80. [19] Koenig C, Tick L, Hanna B. Analyses of the flash track DNA probe and UTIscreen bioluminescence tests for bacteriuria. J Clin Microbiol 1992;30:342–5. [20] Marschal M, Wienke M, Hoering S, Autenrieth IB, Frick JS. Evaluation of 3 different rapid automated systems for diagnosis of urinary tract infections. Diagn Microbiol Infect Dis 2012;72:125–30. [21] Milagro A, Moles B, Seoane A, Navascués J, Blasco M, García-Moya JB. UTIscreen versus UROQUICK: two semiautomatic systems for bacteriuria detection. Enferm Infecc Microbiol Clin 1999;17:398–400. [22] Murray P, Smith T, McKinney R. Clinical evaluation of three urine screening tests. J Clin Microbiol 1987;25:467–70. [23] Koken T, Aktepe O, Sertesr M, Samli M, Kahraman A, Dogan N. Determination of cutoff values for leukocytes and bacteria for urine flow cytometer (UF-100) in urinary tract infection. Int Urol Nephrol 2002;34:175–8. [24] Gutiérrez-Fernández J, Lara A, Bautista MF, et al. Performance of the Sysmex UF1000i system in screening for significant bacteriuria before quantitative culture of aerobic/ facultative fast-growth bacteria in a reference hospital. J Appl Microbiol 2012;113: 609–14. [25] Falbo R, Sala R, Signorelli S, Venturini N, Sisnorini S, Brambilla P. Bacteriuria screening by automated whole-field-image-based microscopy reduces the number of necessary urine cultures. J Clin Microbiol 2012;50:1427–9. [26] Hayami H, Takahashi S, Ishikawa K, et al. Nationwide surveillance of bacterial pathogens from patients with acute uncomplicated cystitis conducted by the Japanese surveillance committee during 2009 and 2010: antimicrobial susceptibility of Escherichia coli and Staphylococcus saprophyticus. J Infect Chemother 2013;19: 393–403. [27] Neuzillet Y, Naber KG, Schito G, Gualco L, Botto H. French results of the ARESC study: clinical aspects and epidemiology of antimicrobial resistance in female patients with cystitis. Implications for empiric therapy. Med Mal Infect 2012;42:66–75. [28] Parish A, Holliday K. Long-term care acquired urinary tract infections' antibiotic resistance patterns and empiric therapy: a pilot study. Geriatr Nurs 2012;33:473–8. [29] Wada A, Kono M, Kawauchi S, Takagi Y, Morikawa T, Funakoshi K. Rapid discrimination of Gram-positive and Gram-negative bacteria in liquid samples by using NaOHsodium dodecyl sulfate solution and flow cytometry. PLoS ONE 2012;7:e47093. http://dx.doi.org/10.1371/journal.pone.0047093.