- Email: [email protected]
Relationship between conventional culture and flow cytometry for the diagnosis of urinary tract infection
Accepted Manuscript Relationship between conventional culture and flow cytometry for the diagnosis of urinary tract infection
Marta García-Coca, Ignacio Gadea, Jaime Esteban PII: DOI: Reference:
S0167-7012(17)30071-4 doi: 10.1016/j.mimet.2017.03.010 MIMET 5137
To appear in:
Journal of Microbiological Methods
Received date: Revised date: Accepted date:
23 January 2017 17 March 2017 17 March 2017
Please cite this article as: Marta García-Coca, Ignacio Gadea, Jaime Esteban , Relationship between conventional culture and flow cytometry for the diagnosis of urinary tract infection. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Mimet(2017), doi: 10.1016/j.mimet.2017.03.010
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
REVISED Title: RELATIONSHIP BETWEEN CONVENTIONAL CULTURE AND FLOW CYTOMETRY FOR THE DIAGNOSIS OF URINARY TRACT INFECTION.
Authors names and affiliations: Marta García-Coca, Ignacio Gadea, Jaime Esteban
PT
Department of Clinical Microbiology, IIS-Fundación Jiménez Díaz, UAM. Av. Reyes Católicos 2. 28040-Madrid, Spain.Spain.
SC
RI
Marta García-Coca, mail-to: [email protected] Ignacio Gadea, mail-to: [email protected] Jaime Esteban, mail-to: [email protected]
MA
NU
Corresponding author: Jaime Esteban MD, PhD. Department of Clinical Microbiology. IIS-Fundación Jiménez Díaz. Av. Reyes Católicos 2. 28040-Madrid, Spain. mail-to: [email protected]
KEYWORDS: Flow cytometry; Urine culture; Urinary tract infection; Sysmex
PT E
D
ABBREVIATIONS: FCA, flow cytometry analyzer.
CE
HIGHLIGHTS: Screening of urines with flow cytometry analyzers reduces culture workload. Particular flow cytometry parameters can be correlated with main UTI pathogens. Flow cytometry may be a potential tool to predict the etiology of UTI.
AC
DISCLOSER: All authors disclose no conflict of interest. All authors have contributed to study design, sample analysis, manuscript conception and final approval of the article. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Sysmex España S.L. financially supported the study providing reagents and UF-1000i equipment.
ACKNOWLEDGEMENTS: We thank Sysmex España for their technical support.
1
ACCEPTED MANUSCRIPT ABSTRACT BACKGROUND: Urine culture is the gold standard for the diagnosis of urinary tract infections (UTI). The use of flow cytometry analyzers (FCA) prior to culture allows for the quantification and recognition of cell components in urine to be automated and makes it possible to relate these data to the urine pathogens subsequently identified in
PT
cultures. METHODS: Urine samples were assessed with the Sysmex UF-1000i analyzer. Those
RI
that met the criteria for culture (>25 leukocytes/μL or >385 bacteria/μL) were subjected
SC
to quantitative urine culture on chromogenic agar. Counts of red blood cells (RBC),
identified in cultures were evaluated.
NU
white blood cells (WBC), epithelial cells (EC), and the kind of microorganisms
MA
RESULTS: A total of 17483 samples were processed by FCA. Of these, 9057 met the criteria for culture. Urine cultures were reduced by 48.2%. The most common urine
D
pathogen was Escherichia coli (60.3%). Negative urine cultures were significantly
PT E
(p<0.001) associated with a lower WBC count than urine with E. coli, Klebsiella spp. and Proteus spp., but urine with Enterococcus spp. had a lower WBC than negative urine. Contaminated urine had a significantly (p<0.001) lower WBC than urine with E.
CE
coli, Klebsiella spp. and Proteus spp, but no differences were found for Enterococcus
AC
spp. (p=0.729). Negative urine cultures had significantly (p<0.05) higher EC than all positive urine samples. Contaminated urine was associated (p<0.001) with higher EC than cultures with E. coli and Klebsiella spp., in comparison with cultures with Enterococcus spp. (p=0.091) and Proteus spp. (p=0.251). CONCLUSION: The use of the Sysmex UF-1000i flow cytometer for screening urine samples allows for a reduction in the number of urine cultures. WBC values correlate well with the main urine pathogens related to UTI. The results observed for Enterococcus spp. suggest a low impact of these pathogens as a cause of UTI. 2
ACCEPTED MANUSCRIPT 1. INTRODUCTION Urinary tract infection (UTI) is one of the most common infections in humans (Foxman, 2010), with urine being the most frequent sample that is submitted for culture examination in clinical microbiology laboratories (Muñoz-Algarra et al., 2013). UTI typically exhibits bacteriuria and pyuria, and urine culture is the gold standard for
PT
etiological diagnosis (Giesen et al., 2013; Medina-Bombardó and Jover-Palmer, 2011; Schmiemann et al., 2010), despite its long turnaround time (24 to 48 hours). The
RI
screening of urine by flow cytometer analyzers (FCA) can lead to a reduction in urine
SC
cultures and faster reporting of negative results, avoiding unnecessary antibiotic treatment (de Frutos-Serna et al., 2014; Gómez-camarasa et al., 2015). Furthermore, the
NU
examination of urine with FCA prior to culture can improve the diagnosis by correlating
MA
FCA parameters, such as the presence of white blood cells (WBC), red blood cells (RBC) and epithelial cells (EC), with the kind of microorganisms isolated in urine
D
samples from patients with a suspected UTI (Monsen and Rydén, 2015). The Sysmex
PT E
UF-1000i analyzer is a flow cytometer analyzer equipped with a laser and two analysis channels, one for bacteria and another one for other urine forming elements. Using a fluorescent dye that stains DNA, the analyzer counts particles based on internal and
CE
external structures, size and fluorescence (Giesen et al., 2013; Moshaver et al., 2016).
AC
Use of the Sysmex UF-1000i analyzer for screening urine to rule out suspected UTI is feasible according to several previous studies (Broeren et al., 2011; Kadkhoda et al., 2011; Le et al., 2016; Marschal et al., 2012; Pieretti et al., 2010).
The aim of this study was to investigate whether the automated flow cytometer parameters used for screening urine to rule out suspected UTI in current clinical practice correlate with the cultural isolation of the most significant bacterial urine pathogens.
3
ACCEPTED MANUSCRIPT 2. MATERIALS AND METHODS A retrospective study was conducted with urine samples assessed at the Microbiology Laboratory between June and September of 2015. Urines were collected from primary health care, outpatients, emergency and hospitalized patients. Samples from children were collected using pediatric bags and samples from adults were obtained by clean-catch midstream
PT
technique, using sterile wide-rim containers (approximately 10 milliliters of sample). 1.2 milliliters of urines samples were assessed with the Sysmex UF-1000i analyzer in the
RI
same day of arrival at Laboratory. Only those that met the criteria for culture (>25
SC
WBC/μL or >385 bacteria/μL) according to the manufacturer’s recommendations were subjected to a quantitative aerobic culture on chromogenic agar plates (CPS agar;
NU
bioMérieux, Marcy- l’Etoile, France) and incubated at 37ºC for 16-24 hours. Culture criteria and satisfactory accuracy were confirmed prior to the study using a preliminary
MA
set of urine samples (data not shown). Samples were excluded from FCA analysis if excessive turbidity or blood was noted on visual inspection. Culture results were
D
reported semi-quantitatively and classified in four categories as follows: negative, <104
PT E
CFU/mL, 104-105 CFU/mL, >105 CFU/mL. Urine samples were defined as contaminated if more than two different pathogen species were detected in cultures. The
CE
diversity of microorganisms in contaminated urines was not documented. The microorganisms were identified by morphology and pigment in the CPS agar or using
AC
MALDI-TOF methodology (Vitek MS©, bioMérieux, Marcy-l’Etoile, France). Patient demographics (age and gender), source of the urine sample, culture outcomes (CFU/mL) and flow cytometry parameters such as RBC, WBC and EC values were recorded and evaluated. Data were analyzed using NCSS10 software 2015 for descriptive analysis. Dunn tests (R1.3.2 software) were conducted for comparisons between groups. Differences between groups were analyzed using the Kruskal-Wallis test. In the event of significant 4
ACCEPTED MANUSCRIPT differences, multiple Dunn test comparisons were performed, adjusting the p-value by the Benjamini-Hochberg method. The diagnostic performance with regard to the FCA variables bacteriuria and leukocyturia detected with the Sysmex UF-1000i as predictors of urine culture results was evaluated using Receiver Operating Characteristic (ROC) curves.
PT
The BACT-morph software of Sysmex UF-1000i discriminates bacterial morphology
RI
and classifies bacteria as rods or cocci/mixed. Briefly, a fluorescent dye of nucleic acids added to the urine sample is detected in the bacteria channel. The forward-scattered
SC
fluorescence discriminates particle size while the side-scattered fluorescence indicates
NU
particle surface and complexity. Subsequently, the Sysmex UF-1000i analyzer classifies bacteria according to their fluorescence emission as rods or cocci/mixed. The
MA
performance of Sysmex UF-1000i to identify the presence of an infection caused by bacilli was evaluated in terms of sensitivity (Se), specificity (Sp), positive predictive
PT E
D
value (PPV) and negative predictive value (NPV).
3. RESULTS
CE
A total of 17,483 urine samples were assessed with the Sysmex UF-1000i flow
AC
cytometer analyzer. Of these, 9057 (51.8%) met the criteria for culture examination. The reduction in urine cultures resulting from use of the flow cytometry screening was 48.2%. The median age of the patients was 47 years (P25-P75: 33.0-73.0) and most of them were women [85.3% (95% confidence interval: 84.6-86.0)]. Most samples originated in primary health care [77.7% (76.8-78.5)], and fewer from outpatients [15.7% (15.0-16.5)], emergency [4.1% (3.7-4.5)] and hospitalized patients [2.6% (2.32.9)]. In urine culture assessments we found 3725 [41.1%; (40.1-42.2)] negative cultures, 2146 [23.7% (22.8-24.6)] contaminated urines and 3186 [35.2% (34.2-36.2)] 5
ACCEPTED MANUSCRIPT urines with bacterial growth. The most common pathogens were Escherichia coli [60.3% (58.6-62.0)], Klebsiella spp. [14.9% (13.7-16.1)] and Enterococcus spp. [12.0% (10.9-13.3)]. Demographic data, urine sample characteristics and culture results are shown in Table 1. Significant differences (p<0.05) in WBC, RBC, EC and bacteria counts were observed
PT
in certain culture results (Table 2 and Table 3). Negative urine cultures showed
RI
significant (p<0.05) differences for all forming element counts in comparison with all other groups (Table 3). In particular, they yielded WBC counts (28.9 cells/μL) lower
SC
than E. coli, Klebsiella spp. and Proteus spp. (55.9, 55.1 and 54.4 cells/μL,
NU
respectively), but higher than Enterococcus spp. (18.9 cells/μL) and contaminated urine (15.6 cells/μL), as well as higher EC counts, with the exception of Enterococcus spp.
MA
(p=a) and Proteus spp. Cultures with E. coli and Klebsiella spp. growth did not show significantly different WBC, RBC, EC or bacteria counts. E. coli and Proteus spp.
D
showed significant (p<0.001) differences only in EC counts (4.6 cells/µL versus 7.4
PT E
cells/µL, respectively). E. coli and Enterococcus spp. showed significantly different values for all counts, with higher WBC counts but lower RBC and EC counts in E.coli
CE
cultures (12.6 cells/µL and 4.6 cells/µL versus 14.0 cells/µL and 4.2 cells/µL, respectively). Contaminated urine samples had significantly (p<0.001) lower WBC
AC
counts than those with E. coli, Klebsiella spp. and Proteus spp., but no differences (p = 0.248) were found with Enterococcus spp. (15.6 cells/µL and 18.9 cells/µL, respectively). No differences in EC count were observed between contaminated urine samples and samples with Enterococcus spp. (p = 0.091) or Proteus spp. (p = 0.251). Performance in the diagnosis of gram-negative bacilli with Sysmex UF-1000i using the BACT-morph software showed a sensitivity of 81.4% (79.8-86.9), specificity of 48.1%
6
ACCEPTED MANUSCRIPT (43.4-52.8), positive predictive value of 90.4% (89.1-91.5) and negative predictive value of 30% (26.7-33.6). In the total sample analysis, the diagnostic performance of bacteriuria and leukocyturia as predictors of urine culture results yielded an area under the curve (AUC) value of 0.782 (0.772-0.792) for bacteria count and 0.605 (0.593-0.618) for WBC count (Figure
PT
1A). Excluding contaminated samples, the AUC increased to 0.844 (0.834-0.853) for
RI
bacteria, but decreased to 0.583 (0.570-0.597) for WBC (Figure 1B). When the analysis sample was restricted to positive cultures, samples with over 105 CFU/mL of bacteria
SC
(Figure 1C), the AUC values were 0.798 (0.788-0.808) and 0.607 (0.594-0.619) for
MA
NU
bacteria and WBC, respectively.
4. DISCUSSION
As demonstrated by several other studies, the screening of urine samples using FCA
PT E
D
makes it possible to rule out suspected UTI, reducing the number of samples sent for culture, workload and costs (Broeren et al., 2011; Kadkhoda et al., 2011; Le et al., 2016; Marschal et al., 2012; Monsen and Rydén, 2015; Pieretti et al., 2010). Furthermore,
CE
turnaround time can also be reduced to the extent that negative results in urine cultures
AC
can be reported on the same day and unnecessary antibiotic prescriptions can be avoided (Giesen et al., 2013; Le et al., 2016). Few studies, however, have focused on the feasibility of FCA being used to predict urine pathogens prior to culture. In this study we have evaluated the correlation of urine culture results (such as urine pathogen types, absence of bacterial growth and contaminated specimens) with the WBC, RBC and EC counts assessed with the Sysmex UF-1000i analyzer used in current clinical practice to rule out UTI in our hospital.
7
ACCEPTED MANUSCRIPT FCA performance was confirmed prior to the etiological characterization of the UTI. The ROC analyses yielded results similar to those of other previously reported studies that demonstrated that FCA detects bacteriuria more accurately than leukocyturia. Nevertheless, the combination of both leukocyte and bacteria count is an effective predictor of UTI (Broeren et al., 2011; Pieretti et al., 2010).
PT
The profile of urine pathogens observed in the present study correlates with recent UTI
RI
epidemiology reports (Ejrnæs et al., 2011; Foxman, 2010; Monsen and Rydén, 2015; Ronald, 2003), with E.coli being the main urine pathogen in cases of outpatient UTI,
SC
followed by two commonly described enterobacteria and Enterococcus spp. According
NU
to FCA parameters, the higher WBC counts in urine samples with gram-negative and gram-positive pathogens (except for enterococci) in comparison with negative samples
MA
indicate good correlation for the etiological diagnosis of UTI between urine culture results and FCA analysis. We found that UTI with E. coli and Klebsiella spp. have
D
similar WBC, RBC, EC and bacterial counts. Similar results were recently reported by
PT E
Monsen et al. (Monsen and Rydén, 2015). In addition, we have observed that the WBC counts in UTI with Enterococcus spp. were significantly lower than in UTI with other
CE
microorganisms, or even than in negative samples. Furthermore, no differences for RBC, WBC, EC and bacterial counts were found between contaminated samples and
AC
urine with Enterococcus spp. growth. Data on Enterococcus spp. suggest that these pathogens are not very significant as a cause of UTI. These results agree with the conclusion of Hooton et al., who reported that acute uncomplicated cystitis is rarely caused solely by enterococci (Hooton et al., 2013). With regard to EC values in contaminated and negative urine samples, the EC count could be a good indicator of contaminated urine, which could prevent false positive results due to sample contamination (de Frutos-Serna et al., 2014). Interestingly, urine samples with Proteus spp. growth did not show a significant difference in EC count compared with 8
ACCEPTED MANUSCRIPT contaminated urine samples. Proteus mirabilis is more strongly associated with human urinary sediment containing epithelial cells rather than other enterobacteria (Eden et al., 1980), which suggests the possibility that many of these pathogens were actually skin contaminations and not real UTI. In addition, the Sysmex UF-1000i analyzer is equipped with the bacterial morphology
PT
(BACT-morph) software, which distinguishes bacilli from cocci/mixed pathogens in
RI
urine and could guide the physician towards the etiology of UTI prior to urine culture results. The diagnostic performance in the diagnosis of gram-negative bacilli using
SC
BACT-morph was acceptable in our settings. Geerts et al. reported a 91% concordance
NU
between BACT-morph and culture in the identification of bacilli. A recent study reported moderate diagnostic accuracy (78.4%) with similar sensitivity and specificity
MA
results (82.4% and 62.5%, respectively), and concluded that the software offers approximate differentiation of pathogenic urinary bacilli (Jiménez-Guerra et al., 2016).
D
Furthermore, Geerts et al. assessed the response to a first-line antibiotic based on the
PT E
foreknowledge of bacterial morphology provided by BACT-morph. They estimated that in a hypothetical scenario with perfect accuracy of BACT-morph performance, the
CE
overall response (bacilli and cocci) would be 79%.
AC
Diagnosing on the basis of symptoms alone will overestimate the presence of UTI by 33%, which can lead to unnecessary prescribing of antibiotics and selection for antibiotic-resistant bacteria (Foxman, 2014; Shepherd and Pottinger, 2013). UTI are often caused by bacilli, and these are generally considered uncomplicated UTI that can be treated easily with a standard antibiotic. In contrast, UTI caused by cocci are more complicated and treatment usually requires a more aggressive type of antibiotic (Shepherd and Pottinger, 2013; Shigemura et al., 2009). Therefore, prediction of bacteria type, as well as bacteria count, may guide the physician in the choice of 9
ACCEPTED MANUSCRIPT antibiotic. Once culture results are available, the antibiotic prescription can be switched, if needed. FCA parameters and tools such as the BACT-morph software may help, to the extent that the urine pathogens can be classified based on FCA characteristics and identified prior to culture. This study has some limitations. In this study, screening with FCA in current laboratory
PT
practice entails particular culture criteria that exclude negative urine samples and urine
RI
samples with low cell counts, although these are included in other studies (de FrutosSerna et al., 2014; Monsen and Rydén, 2015). This creates an unavoidable bias, which
SC
could explain why the results we obtained for negative urine samples in some cases
NU
show high cell counts, which could be related to pathologies other than a UTI. Secondly, microorganisms that are not cultivable on standard agars, such as N.
MA
gonorrhoeae and Chlamydia, are overlooked by the study design. This also adds a bias that would reduce the accuracy of the FCA parameters as a tool to predict UTI etiology.
D
Further investigations of urines from symptomatic patients with negative cultures but
PT E
significant FCA parameters should be addressed. 5. CONCLUSIONS
CE
In our current clinical practice, particular counts for flow cytometry parameters such as
AC
WBC, RBC and EC are associated with specific types of urine pathogens. The Sysmex UF-1000i analyzer can be a useful tool for the diagnosis of UTI etiology and one that, in combination with conventional methodologies, assists clinicians in choosing the most appropriate antibiotic treatment. Further prospective studies are needed to evaluate the performance of antibiotics chosen based on the FCA characteristics of urine samples in suspected UTI.
10
ACCEPTED MANUSCRIPT 6. REFERENCES Broeren, M. a C., Bahçeci, S., Vader, H.L., Arents, N.L.A., 2011. Screening for urinary tract infection with the sysmex UF-1000i urine flow cytometer. J. Clin. Microbiol. 49, 1025–1029. doi:10.1128/JCM.01669-10 de Frutos-Serna, M. De, Asensio-calle, M.L., Haro-pérez, A.M., María, A., Castro, B., Nieves, M., Iglesias-garcía, J., 2014. Evaluación del citómetro UF-1000i como método de cribado en el diagnóstico de infecciones del tracto urinario 32, 147–151.
PT
Eden, C.S., Larsson, P., Lomberg, H., 1980. Attachment of Proteus mirabilis to human urinary sediment epithelial cells in vitro is different from that of Escherichia coli.
RI
Infect. Immun. 27, 804–807.
SC
Ejrnæs, K., Stegger, M., Reisner, A., Ferry, S., Monsen, T., Holm, S.E., Lundgren, B., Frimodt-Møller, N., 2011. Characteristics of Escherichia coli causing persistence or relapse of urinary tract infections: Phylogenetic groups, virulence factors and
NU
biofilm formation. Virulence 2, 528–537. doi:10.4161/viru.2.6.18189 Foxman, B., 2010. The epidemiology of urinary tract infection. Nat. Rev. Urol. 7, 653–
MA
60. doi:10.1038/nrurol.2010.190
Foxman, B., 2014. Urinary tract infection syndromes. Occurrence, recurrence, bacteriology, risk factors, and disease burden. Infect. Dis. Clin. North Am. 28, 1–
D
13. doi:10.1016/j.idc.2013.09.003
PT E
Geerts, N., Jansz, A.R., Boonen, K.J.M., Wijn, R.P.W.F., Koldewijn, E.L., Boer, A.K., Scharnhorst, V., 2015. Urine flow cytometry can rule out urinary tract infection, but cannot identify bacterial morphologies correctly. Clin. Chim. Acta 448, 86–90.
CE
doi:10.1016/j.cca.2015.06.020
Giesen, C.D., Greeno, A.M., Thompson, K.A., Patel, R., Jenkins, S.M., Lieske, J.C.,
AC
2013. Performance of flow cytometry to screen urine for bacteria and white blood cells prior to urine culture. Clin. Biochem. 46, 810–813. doi:10.1016/j.clinbiochem.2013.03.005 Gómez-Camarasa, C., Liébana-Martos, C., Navarro-Marí, J.M., 2015. Original Detección de uropatógenos inusuales durante un periodo de 3 años en un hospital regional 28, 86–91. Hooton, T.M., Roberts, P.L., Cox, M.E., Stapleton, A.E., 2013. Voided Midstream Urine Culture and Acute Cystitis in Premenopausal Women 369, 1883–1891. doi:10.1056/NEJMoa1302186.Voided Jiménez-Guerra, G., Heras-Cañas, V., Valera-Arcas, M.D., Rodríguez-Grangér, J., 11
ACCEPTED MANUSCRIPT Navarro, J.M., Gutiérrez-Fernández, J., 2016. Comparison between urine culture profile and morphology classification using fluorescence parameters of the sysmex uf-1000i urine flow cytometer. J. Appl. Microbiol. doi:10.1111/jam.13354 Kadkhoda, K., Manickam, K., DeGagne, P., Sokolowski, P., Pang, P., Kontzie, N., Alfa, M., 2011. UF-1000i flow cytometry is an effective screening method for urine specimens. Diagn. Microbiol. Infect. Dis. 69, 130–136. doi:10.1016/j.diagmicrobio.2010.09.013
PT
Le, Z., Li, F., Fei, C., Ye, A., Xie, X., Zhang, J., 2016. Performance of the Sysmex UF1000i urine analyser in the rapid diagnosis of urinary tract infections in
RI
hospitalized patients. J. Infect. Chemother. doi:10.1016/j.jiac.2016.02.009 Marschal, M., Wienke, M., Hoering, S., Autenrieth, I.B., Frick, J.S., 2012. Evaluation
Diagn. Microbiol. Infect. Dis. 72, 125–130.
NU
doi:10.1016/j.diagmicrobio.2011.10.001
SC
of 3 different rapid automated systems for diagnosis of urinary tract infections.
Medina-Bombardó, D., Jover-Palmer, A., 2011. Does clinical examination aid in the
MA
diagnosis of urinary tract infections in women? A systematic review and metaanalysis. BMC Fam. Pract. 12, 111.
Monsen, T., Rydén, P., 2015. Flow cytometry analysis using sysmex UF-1000i
D
classifies uropathogens based on bacterial, leukocyte, and erythrocyte counts in
PT E
urine specimens among patients with urinary tract infections. J. Clin. Microbiol. 53, 539–545. doi:10.1128/JCM.01974-14 Moshaver, B., de Boer, F., van Egmond-Kreileman, H., Kramer, E., Stegeman, C.,
CE
Groeneveld, P., 2016. Fast and accurate prediction of positive and negative urine cultures by flow cytometry. BMC Infect. Dis. 16, 211. doi:10.1186/s12879-016-
AC
1557-4
Muñoz-Algarra, M., Martínez-Ruiz, R., Orden-Martínez, B., 2013. Evaluación del sistema automatizado UF-1000i en el diagnóstico de infección urinaria. Enferm. Infecc. Microbiol. Clin. 31, 29–31. Pieretti, B., Brunati, P., Pini, B., Colzani, C., Congedo, P., Rocchi, M., Terramocci, R., 2010. Diagnosis of bacteriuria and leukocyturia by automated flow cytometry compared with urine culture. J. Clin. Microbiol. 48, 3990–3996. doi:10.1128/JCM.00975-10 Ronald, A., 2003. The etiology of urinary tract infection. Traditional and emerging pathogens. Dis Mon. 2003; 49:71–82. Schmiemann, G., Kniehl, E., Gebhardt, K., Matejczyk, M.M., Hummers-Pradier, E., 12
ACCEPTED MANUSCRIPT 2010. The diagnosis of urinary tract infection: a systematic review. Dtsch Arztebl Int 107, 361–367. doi:10.3238/arztebl.2010.0361 Shepherd, A.K., Pottinger, P.S., 2013. Management of Urinary Tract Infections in the Era of Increasing Antimicrobial Resistance. Med. Clin. North Am. 97, 737–757. doi:10.1016/j.mcna.2013.03.006 Shigemura, K., Arakawa, S., Tanaka, K., Fujisawa, M., 2009. Clinical investigation of isolated bacteria from urinary tracts of hospitalized patients and their
PT
susceptibilities to antibiotics. J. Infect. Chemother. 15, 18–22. doi:10.1007/s10156-
AC
CE
PT E
D
MA
NU
SC
RI
008-0652-x
13
ACCEPTED MANUSCRIPT Table 1. Demographic characteristics and urine culture data. n
N = 9057
% (95%CI)
Age (years)* Patients
47 (33.0 - 73.0) women
7726
85.3 (84.6 - 86.0)
men
1331
14.7 (14.0 – 15.4)
Primary health care
7034
77.7 (76.8 - 78.5)
Outpatients
1424
Emergency
368
Hospitalized patients
231
Culture results
2.6 (2.3 - 2.9)
41.1 (1 - 42.2)
2146
23.7 (22.8 - 24.6)
3186
35.2 (34.2 - 36.2)
Escherichia coli
1922
60.3 (58.6 – 62.0)
Klebsiella spp.
473
14.9 (13.7 - 16.1)
381
12.0 (10.9 - 13.3)
Proteus spp.
132
4.1 (3.5 - 4.9)
Other gram-negative bacilli
178
5.6 (4.8 - 6.4)
100
3.1 (2.6 - 3.8)
Contaminated
D
MA
Bacterial growth
Enterococcus spp.
CE
PT E
Urine pathogens
4.1 (3.7 - 4.5)
3725
NU
Negative
15.7 (15.0 - 16.5)
RI
SC
Urine sources
PT
Gender
Other microorganisms
AC
95% confidence interval of relative frequency (95%CI). *Median and interquartile range (IQR).
14
ACCEPTED MANUSCRIPT Table 2. Forming element counts in urine samples assessed with FCA organized by culture results. RBC (cells/µL)
EC (cells/µL)
E. coli (n=1922)
55.9 (0.2 - 33,243.2)
12.6 (0.1 - 6361.2)
4.6 (0.0 - 489.3)
Klebsiella spp. (n=473)
55.1 (0.4 - 32,777.6)
12.9 (0.5 - 1977.1)
4.2 (0.0 - 110.6)
Enterococcus spp. (n=381)
18.9 (0.2 - 14,869.8)
14,0 (0.9 - 515.0)
7.6 (0.0 - 157.0)
54.4 (1.0 - 1812.0)
11.3 (1.0 - 1962.9)
7.4 (0.0 - 129.0)
Other gram-negative bacilli (n=178) Other microorganisms (n=100)
43.6 (0.2 - 25,240.5)
RI
SC
Proteus spp. (n=132)
PT
WBC (cells/µL)
12.3 (0.5 - 755.2)
NU
N = 9057
4.4 (0.0 - 80.4)
20.8 (2.0 - 2106.7)
5.65 (0.1 - 259.2)
Contaminated urine (n=2146)
15.6 (0.0 - 14,234.1)
15.3 (0.2 - 37,118.2)
8.8 (0.0 - 519.0)
Negative urine (n=3725)
28.9 (0.1 - 32,621.6)
16.4 (0.2 - 37,593.2)
13.7 (0.0 - 962.7)
D
MA
70.9 (0.3 - 2799.7)
AC
CE
PT E
Data shown as median values and range. WBC, white blood cells; RBC, red blood cells; EC, epithelial cells
15
ACCEPTED MANUSCRIPT Table 3. Pair-wise comparison of urine culture results. Data shown as corrected pvalues (Benjamini-Hochberg method) with significance for p<0.05.
E. coli
Klebsiella
Enterococcus
Proteus
Other
spp
spp
spp
Gram-
E. coli
<0.001
Klebsiella spp
<0.001
0.371
Enterococcus spp
0.004
0.067
0.170
Proteus spp.
0.001
0.316
0.276
0.103
Other Gram-
<0.001
0.216
0.190
0.053
Others
0.030
<0.001
<0.001
0.002
Contamination
0.007
<0.001
0.002
0.095
Negative
E. coli
<0.001
<0.001
0.010
0.002
Klebsiella
Enterococcus
Proteus
Other
spp
spp
spp
Gram-
<0.001
Klebsiella spp
<0.001
0.239
Enterococcus spp
<0.001
<0.001
<0.001
Proteus spp.
<0.001
0.237
0.145
<0.001
Other Gram-
0.013
0.029
0.117
<0.001
0.028
Others
<0.001
0.343
0.235
<0.001
0.411
0.067
Contamination
<0.001
<0.001
<0.001
0.248
<0.001
<0.001
MA
NU
E. coli
Enterococcus
Proteus
Other
spp
spp
spp
Gram-
E. coli
E. coli
<0.001
D
Klebsiella
Negative
Klebsiella spp
<0.001
0.230
Enterococcus spp
<0.001
<0.001
<0.001
Proteus spp.
<0.001
<0.001
<0.001
0.469
Other Gram-
<0.001
0.310
0.499
<0.001
<0.001
<0.001
0.058
0.036
0.048
0.062
0.056
<0.001
<0.001
<0.001
0.091
0.251
<0.001
Others
<0.001 Others
0.004
AC
Contamination
CE
Others
PT E
EC
0.005
SC
WBC
Others
0.426
PT
Negative
RI
RBC
16
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Figure 1. Diagnostic accuracy with regard to pathogen growth in urine cultures according to bacteria and WBC count in flow cytometry. ROC curve analysis with different subsets of urine culture results. (A) All urine samples; (B) exclusion of contaminated urine samples; (C) only positive urine cultures with bacterial counts over 105 CFU/mL. Blue line (bacterial) and grey line (WBC).
17
Marta García-Coca, Ignacio Gadea, Jaime Esteban PII: DOI: Reference:
S0167-7012(17)30071-4 doi: 10.1016/j.mimet.2017.03.010 MIMET 5137
To appear in:
Journal of Microbiological Methods
Received date: Revised date: Accepted date:
23 January 2017 17 March 2017 17 March 2017
Please cite this article as: Marta García-Coca, Ignacio Gadea, Jaime Esteban , Relationship between conventional culture and flow cytometry for the diagnosis of urinary tract infection. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Mimet(2017), doi: 10.1016/j.mimet.2017.03.010
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
REVISED Title: RELATIONSHIP BETWEEN CONVENTIONAL CULTURE AND FLOW CYTOMETRY FOR THE DIAGNOSIS OF URINARY TRACT INFECTION.
Authors names and affiliations: Marta García-Coca, Ignacio Gadea, Jaime Esteban
PT
Department of Clinical Microbiology, IIS-Fundación Jiménez Díaz, UAM. Av. Reyes Católicos 2. 28040-Madrid, Spain.Spain.
SC
RI
Marta García-Coca, mail-to: [email protected] Ignacio Gadea, mail-to: [email protected] Jaime Esteban, mail-to: [email protected]
MA
NU
Corresponding author: Jaime Esteban MD, PhD. Department of Clinical Microbiology. IIS-Fundación Jiménez Díaz. Av. Reyes Católicos 2. 28040-Madrid, Spain. mail-to: [email protected]
KEYWORDS: Flow cytometry; Urine culture; Urinary tract infection; Sysmex
PT E
D
ABBREVIATIONS: FCA, flow cytometry analyzer.
CE
HIGHLIGHTS: Screening of urines with flow cytometry analyzers reduces culture workload. Particular flow cytometry parameters can be correlated with main UTI pathogens. Flow cytometry may be a potential tool to predict the etiology of UTI.
AC
DISCLOSER: All authors disclose no conflict of interest. All authors have contributed to study design, sample analysis, manuscript conception and final approval of the article. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Sysmex España S.L. financially supported the study providing reagents and UF-1000i equipment.
ACKNOWLEDGEMENTS: We thank Sysmex España for their technical support.
1
ACCEPTED MANUSCRIPT ABSTRACT BACKGROUND: Urine culture is the gold standard for the diagnosis of urinary tract infections (UTI). The use of flow cytometry analyzers (FCA) prior to culture allows for the quantification and recognition of cell components in urine to be automated and makes it possible to relate these data to the urine pathogens subsequently identified in
PT
cultures. METHODS: Urine samples were assessed with the Sysmex UF-1000i analyzer. Those
RI
that met the criteria for culture (>25 leukocytes/μL or >385 bacteria/μL) were subjected
SC
to quantitative urine culture on chromogenic agar. Counts of red blood cells (RBC),
identified in cultures were evaluated.
NU
white blood cells (WBC), epithelial cells (EC), and the kind of microorganisms
MA
RESULTS: A total of 17483 samples were processed by FCA. Of these, 9057 met the criteria for culture. Urine cultures were reduced by 48.2%. The most common urine
D
pathogen was Escherichia coli (60.3%). Negative urine cultures were significantly
PT E
(p<0.001) associated with a lower WBC count than urine with E. coli, Klebsiella spp. and Proteus spp., but urine with Enterococcus spp. had a lower WBC than negative urine. Contaminated urine had a significantly (p<0.001) lower WBC than urine with E.
CE
coli, Klebsiella spp. and Proteus spp, but no differences were found for Enterococcus
AC
spp. (p=0.729). Negative urine cultures had significantly (p<0.05) higher EC than all positive urine samples. Contaminated urine was associated (p<0.001) with higher EC than cultures with E. coli and Klebsiella spp., in comparison with cultures with Enterococcus spp. (p=0.091) and Proteus spp. (p=0.251). CONCLUSION: The use of the Sysmex UF-1000i flow cytometer for screening urine samples allows for a reduction in the number of urine cultures. WBC values correlate well with the main urine pathogens related to UTI. The results observed for Enterococcus spp. suggest a low impact of these pathogens as a cause of UTI. 2
ACCEPTED MANUSCRIPT 1. INTRODUCTION Urinary tract infection (UTI) is one of the most common infections in humans (Foxman, 2010), with urine being the most frequent sample that is submitted for culture examination in clinical microbiology laboratories (Muñoz-Algarra et al., 2013). UTI typically exhibits bacteriuria and pyuria, and urine culture is the gold standard for
PT
etiological diagnosis (Giesen et al., 2013; Medina-Bombardó and Jover-Palmer, 2011; Schmiemann et al., 2010), despite its long turnaround time (24 to 48 hours). The
RI
screening of urine by flow cytometer analyzers (FCA) can lead to a reduction in urine
SC
cultures and faster reporting of negative results, avoiding unnecessary antibiotic treatment (de Frutos-Serna et al., 2014; Gómez-camarasa et al., 2015). Furthermore, the
NU
examination of urine with FCA prior to culture can improve the diagnosis by correlating
MA
FCA parameters, such as the presence of white blood cells (WBC), red blood cells (RBC) and epithelial cells (EC), with the kind of microorganisms isolated in urine
D
samples from patients with a suspected UTI (Monsen and Rydén, 2015). The Sysmex
PT E
UF-1000i analyzer is a flow cytometer analyzer equipped with a laser and two analysis channels, one for bacteria and another one for other urine forming elements. Using a fluorescent dye that stains DNA, the analyzer counts particles based on internal and
CE
external structures, size and fluorescence (Giesen et al., 2013; Moshaver et al., 2016).
AC
Use of the Sysmex UF-1000i analyzer for screening urine to rule out suspected UTI is feasible according to several previous studies (Broeren et al., 2011; Kadkhoda et al., 2011; Le et al., 2016; Marschal et al., 2012; Pieretti et al., 2010).
The aim of this study was to investigate whether the automated flow cytometer parameters used for screening urine to rule out suspected UTI in current clinical practice correlate with the cultural isolation of the most significant bacterial urine pathogens.
3
ACCEPTED MANUSCRIPT 2. MATERIALS AND METHODS A retrospective study was conducted with urine samples assessed at the Microbiology Laboratory between June and September of 2015. Urines were collected from primary health care, outpatients, emergency and hospitalized patients. Samples from children were collected using pediatric bags and samples from adults were obtained by clean-catch midstream
PT
technique, using sterile wide-rim containers (approximately 10 milliliters of sample). 1.2 milliliters of urines samples were assessed with the Sysmex UF-1000i analyzer in the
RI
same day of arrival at Laboratory. Only those that met the criteria for culture (>25
SC
WBC/μL or >385 bacteria/μL) according to the manufacturer’s recommendations were subjected to a quantitative aerobic culture on chromogenic agar plates (CPS agar;
NU
bioMérieux, Marcy- l’Etoile, France) and incubated at 37ºC for 16-24 hours. Culture criteria and satisfactory accuracy were confirmed prior to the study using a preliminary
MA
set of urine samples (data not shown). Samples were excluded from FCA analysis if excessive turbidity or blood was noted on visual inspection. Culture results were
D
reported semi-quantitatively and classified in four categories as follows: negative, <104
PT E
CFU/mL, 104-105 CFU/mL, >105 CFU/mL. Urine samples were defined as contaminated if more than two different pathogen species were detected in cultures. The
CE
diversity of microorganisms in contaminated urines was not documented. The microorganisms were identified by morphology and pigment in the CPS agar or using
AC
MALDI-TOF methodology (Vitek MS©, bioMérieux, Marcy-l’Etoile, France). Patient demographics (age and gender), source of the urine sample, culture outcomes (CFU/mL) and flow cytometry parameters such as RBC, WBC and EC values were recorded and evaluated. Data were analyzed using NCSS10 software 2015 for descriptive analysis. Dunn tests (R1.3.2 software) were conducted for comparisons between groups. Differences between groups were analyzed using the Kruskal-Wallis test. In the event of significant 4
ACCEPTED MANUSCRIPT differences, multiple Dunn test comparisons were performed, adjusting the p-value by the Benjamini-Hochberg method. The diagnostic performance with regard to the FCA variables bacteriuria and leukocyturia detected with the Sysmex UF-1000i as predictors of urine culture results was evaluated using Receiver Operating Characteristic (ROC) curves.
PT
The BACT-morph software of Sysmex UF-1000i discriminates bacterial morphology
RI
and classifies bacteria as rods or cocci/mixed. Briefly, a fluorescent dye of nucleic acids added to the urine sample is detected in the bacteria channel. The forward-scattered
SC
fluorescence discriminates particle size while the side-scattered fluorescence indicates
NU
particle surface and complexity. Subsequently, the Sysmex UF-1000i analyzer classifies bacteria according to their fluorescence emission as rods or cocci/mixed. The
MA
performance of Sysmex UF-1000i to identify the presence of an infection caused by bacilli was evaluated in terms of sensitivity (Se), specificity (Sp), positive predictive
PT E
D
value (PPV) and negative predictive value (NPV).
3. RESULTS
CE
A total of 17,483 urine samples were assessed with the Sysmex UF-1000i flow
AC
cytometer analyzer. Of these, 9057 (51.8%) met the criteria for culture examination. The reduction in urine cultures resulting from use of the flow cytometry screening was 48.2%. The median age of the patients was 47 years (P25-P75: 33.0-73.0) and most of them were women [85.3% (95% confidence interval: 84.6-86.0)]. Most samples originated in primary health care [77.7% (76.8-78.5)], and fewer from outpatients [15.7% (15.0-16.5)], emergency [4.1% (3.7-4.5)] and hospitalized patients [2.6% (2.32.9)]. In urine culture assessments we found 3725 [41.1%; (40.1-42.2)] negative cultures, 2146 [23.7% (22.8-24.6)] contaminated urines and 3186 [35.2% (34.2-36.2)] 5
ACCEPTED MANUSCRIPT urines with bacterial growth. The most common pathogens were Escherichia coli [60.3% (58.6-62.0)], Klebsiella spp. [14.9% (13.7-16.1)] and Enterococcus spp. [12.0% (10.9-13.3)]. Demographic data, urine sample characteristics and culture results are shown in Table 1. Significant differences (p<0.05) in WBC, RBC, EC and bacteria counts were observed
PT
in certain culture results (Table 2 and Table 3). Negative urine cultures showed
RI
significant (p<0.05) differences for all forming element counts in comparison with all other groups (Table 3). In particular, they yielded WBC counts (28.9 cells/μL) lower
SC
than E. coli, Klebsiella spp. and Proteus spp. (55.9, 55.1 and 54.4 cells/μL,
NU
respectively), but higher than Enterococcus spp. (18.9 cells/μL) and contaminated urine (15.6 cells/μL), as well as higher EC counts, with the exception of Enterococcus spp.
MA
(p=a) and Proteus spp. Cultures with E. coli and Klebsiella spp. growth did not show significantly different WBC, RBC, EC or bacteria counts. E. coli and Proteus spp.
D
showed significant (p<0.001) differences only in EC counts (4.6 cells/µL versus 7.4
PT E
cells/µL, respectively). E. coli and Enterococcus spp. showed significantly different values for all counts, with higher WBC counts but lower RBC and EC counts in E.coli
CE
cultures (12.6 cells/µL and 4.6 cells/µL versus 14.0 cells/µL and 4.2 cells/µL, respectively). Contaminated urine samples had significantly (p<0.001) lower WBC
AC
counts than those with E. coli, Klebsiella spp. and Proteus spp., but no differences (p = 0.248) were found with Enterococcus spp. (15.6 cells/µL and 18.9 cells/µL, respectively). No differences in EC count were observed between contaminated urine samples and samples with Enterococcus spp. (p = 0.091) or Proteus spp. (p = 0.251). Performance in the diagnosis of gram-negative bacilli with Sysmex UF-1000i using the BACT-morph software showed a sensitivity of 81.4% (79.8-86.9), specificity of 48.1%
6
ACCEPTED MANUSCRIPT (43.4-52.8), positive predictive value of 90.4% (89.1-91.5) and negative predictive value of 30% (26.7-33.6). In the total sample analysis, the diagnostic performance of bacteriuria and leukocyturia as predictors of urine culture results yielded an area under the curve (AUC) value of 0.782 (0.772-0.792) for bacteria count and 0.605 (0.593-0.618) for WBC count (Figure
PT
1A). Excluding contaminated samples, the AUC increased to 0.844 (0.834-0.853) for
RI
bacteria, but decreased to 0.583 (0.570-0.597) for WBC (Figure 1B). When the analysis sample was restricted to positive cultures, samples with over 105 CFU/mL of bacteria
SC
(Figure 1C), the AUC values were 0.798 (0.788-0.808) and 0.607 (0.594-0.619) for
MA
NU
bacteria and WBC, respectively.
4. DISCUSSION
As demonstrated by several other studies, the screening of urine samples using FCA
PT E
D
makes it possible to rule out suspected UTI, reducing the number of samples sent for culture, workload and costs (Broeren et al., 2011; Kadkhoda et al., 2011; Le et al., 2016; Marschal et al., 2012; Monsen and Rydén, 2015; Pieretti et al., 2010). Furthermore,
CE
turnaround time can also be reduced to the extent that negative results in urine cultures
AC
can be reported on the same day and unnecessary antibiotic prescriptions can be avoided (Giesen et al., 2013; Le et al., 2016). Few studies, however, have focused on the feasibility of FCA being used to predict urine pathogens prior to culture. In this study we have evaluated the correlation of urine culture results (such as urine pathogen types, absence of bacterial growth and contaminated specimens) with the WBC, RBC and EC counts assessed with the Sysmex UF-1000i analyzer used in current clinical practice to rule out UTI in our hospital.
7
ACCEPTED MANUSCRIPT FCA performance was confirmed prior to the etiological characterization of the UTI. The ROC analyses yielded results similar to those of other previously reported studies that demonstrated that FCA detects bacteriuria more accurately than leukocyturia. Nevertheless, the combination of both leukocyte and bacteria count is an effective predictor of UTI (Broeren et al., 2011; Pieretti et al., 2010).
PT
The profile of urine pathogens observed in the present study correlates with recent UTI
RI
epidemiology reports (Ejrnæs et al., 2011; Foxman, 2010; Monsen and Rydén, 2015; Ronald, 2003), with E.coli being the main urine pathogen in cases of outpatient UTI,
SC
followed by two commonly described enterobacteria and Enterococcus spp. According
NU
to FCA parameters, the higher WBC counts in urine samples with gram-negative and gram-positive pathogens (except for enterococci) in comparison with negative samples
MA
indicate good correlation for the etiological diagnosis of UTI between urine culture results and FCA analysis. We found that UTI with E. coli and Klebsiella spp. have
D
similar WBC, RBC, EC and bacterial counts. Similar results were recently reported by
PT E
Monsen et al. (Monsen and Rydén, 2015). In addition, we have observed that the WBC counts in UTI with Enterococcus spp. were significantly lower than in UTI with other
CE
microorganisms, or even than in negative samples. Furthermore, no differences for RBC, WBC, EC and bacterial counts were found between contaminated samples and
AC
urine with Enterococcus spp. growth. Data on Enterococcus spp. suggest that these pathogens are not very significant as a cause of UTI. These results agree with the conclusion of Hooton et al., who reported that acute uncomplicated cystitis is rarely caused solely by enterococci (Hooton et al., 2013). With regard to EC values in contaminated and negative urine samples, the EC count could be a good indicator of contaminated urine, which could prevent false positive results due to sample contamination (de Frutos-Serna et al., 2014). Interestingly, urine samples with Proteus spp. growth did not show a significant difference in EC count compared with 8
ACCEPTED MANUSCRIPT contaminated urine samples. Proteus mirabilis is more strongly associated with human urinary sediment containing epithelial cells rather than other enterobacteria (Eden et al., 1980), which suggests the possibility that many of these pathogens were actually skin contaminations and not real UTI. In addition, the Sysmex UF-1000i analyzer is equipped with the bacterial morphology
PT
(BACT-morph) software, which distinguishes bacilli from cocci/mixed pathogens in
RI
urine and could guide the physician towards the etiology of UTI prior to urine culture results. The diagnostic performance in the diagnosis of gram-negative bacilli using
SC
BACT-morph was acceptable in our settings. Geerts et al. reported a 91% concordance
NU
between BACT-morph and culture in the identification of bacilli. A recent study reported moderate diagnostic accuracy (78.4%) with similar sensitivity and specificity
MA
results (82.4% and 62.5%, respectively), and concluded that the software offers approximate differentiation of pathogenic urinary bacilli (Jiménez-Guerra et al., 2016).
D
Furthermore, Geerts et al. assessed the response to a first-line antibiotic based on the
PT E
foreknowledge of bacterial morphology provided by BACT-morph. They estimated that in a hypothetical scenario with perfect accuracy of BACT-morph performance, the
CE
overall response (bacilli and cocci) would be 79%.
AC
Diagnosing on the basis of symptoms alone will overestimate the presence of UTI by 33%, which can lead to unnecessary prescribing of antibiotics and selection for antibiotic-resistant bacteria (Foxman, 2014; Shepherd and Pottinger, 2013). UTI are often caused by bacilli, and these are generally considered uncomplicated UTI that can be treated easily with a standard antibiotic. In contrast, UTI caused by cocci are more complicated and treatment usually requires a more aggressive type of antibiotic (Shepherd and Pottinger, 2013; Shigemura et al., 2009). Therefore, prediction of bacteria type, as well as bacteria count, may guide the physician in the choice of 9
ACCEPTED MANUSCRIPT antibiotic. Once culture results are available, the antibiotic prescription can be switched, if needed. FCA parameters and tools such as the BACT-morph software may help, to the extent that the urine pathogens can be classified based on FCA characteristics and identified prior to culture. This study has some limitations. In this study, screening with FCA in current laboratory
PT
practice entails particular culture criteria that exclude negative urine samples and urine
RI
samples with low cell counts, although these are included in other studies (de FrutosSerna et al., 2014; Monsen and Rydén, 2015). This creates an unavoidable bias, which
SC
could explain why the results we obtained for negative urine samples in some cases
NU
show high cell counts, which could be related to pathologies other than a UTI. Secondly, microorganisms that are not cultivable on standard agars, such as N.
MA
gonorrhoeae and Chlamydia, are overlooked by the study design. This also adds a bias that would reduce the accuracy of the FCA parameters as a tool to predict UTI etiology.
D
Further investigations of urines from symptomatic patients with negative cultures but
PT E
significant FCA parameters should be addressed. 5. CONCLUSIONS
CE
In our current clinical practice, particular counts for flow cytometry parameters such as
AC
WBC, RBC and EC are associated with specific types of urine pathogens. The Sysmex UF-1000i analyzer can be a useful tool for the diagnosis of UTI etiology and one that, in combination with conventional methodologies, assists clinicians in choosing the most appropriate antibiotic treatment. Further prospective studies are needed to evaluate the performance of antibiotics chosen based on the FCA characteristics of urine samples in suspected UTI.
10
ACCEPTED MANUSCRIPT 6. REFERENCES Broeren, M. a C., Bahçeci, S., Vader, H.L., Arents, N.L.A., 2011. Screening for urinary tract infection with the sysmex UF-1000i urine flow cytometer. J. Clin. Microbiol. 49, 1025–1029. doi:10.1128/JCM.01669-10 de Frutos-Serna, M. De, Asensio-calle, M.L., Haro-pérez, A.M., María, A., Castro, B., Nieves, M., Iglesias-garcía, J., 2014. Evaluación del citómetro UF-1000i como método de cribado en el diagnóstico de infecciones del tracto urinario 32, 147–151.
PT
Eden, C.S., Larsson, P., Lomberg, H., 1980. Attachment of Proteus mirabilis to human urinary sediment epithelial cells in vitro is different from that of Escherichia coli.
RI
Infect. Immun. 27, 804–807.
SC
Ejrnæs, K., Stegger, M., Reisner, A., Ferry, S., Monsen, T., Holm, S.E., Lundgren, B., Frimodt-Møller, N., 2011. Characteristics of Escherichia coli causing persistence or relapse of urinary tract infections: Phylogenetic groups, virulence factors and
NU
biofilm formation. Virulence 2, 528–537. doi:10.4161/viru.2.6.18189 Foxman, B., 2010. The epidemiology of urinary tract infection. Nat. Rev. Urol. 7, 653–
MA
60. doi:10.1038/nrurol.2010.190
Foxman, B., 2014. Urinary tract infection syndromes. Occurrence, recurrence, bacteriology, risk factors, and disease burden. Infect. Dis. Clin. North Am. 28, 1–
D
13. doi:10.1016/j.idc.2013.09.003
PT E
Geerts, N., Jansz, A.R., Boonen, K.J.M., Wijn, R.P.W.F., Koldewijn, E.L., Boer, A.K., Scharnhorst, V., 2015. Urine flow cytometry can rule out urinary tract infection, but cannot identify bacterial morphologies correctly. Clin. Chim. Acta 448, 86–90.
CE
doi:10.1016/j.cca.2015.06.020
Giesen, C.D., Greeno, A.M., Thompson, K.A., Patel, R., Jenkins, S.M., Lieske, J.C.,
AC
2013. Performance of flow cytometry to screen urine for bacteria and white blood cells prior to urine culture. Clin. Biochem. 46, 810–813. doi:10.1016/j.clinbiochem.2013.03.005 Gómez-Camarasa, C., Liébana-Martos, C., Navarro-Marí, J.M., 2015. Original Detección de uropatógenos inusuales durante un periodo de 3 años en un hospital regional 28, 86–91. Hooton, T.M., Roberts, P.L., Cox, M.E., Stapleton, A.E., 2013. Voided Midstream Urine Culture and Acute Cystitis in Premenopausal Women 369, 1883–1891. doi:10.1056/NEJMoa1302186.Voided Jiménez-Guerra, G., Heras-Cañas, V., Valera-Arcas, M.D., Rodríguez-Grangér, J., 11
ACCEPTED MANUSCRIPT Navarro, J.M., Gutiérrez-Fernández, J., 2016. Comparison between urine culture profile and morphology classification using fluorescence parameters of the sysmex uf-1000i urine flow cytometer. J. Appl. Microbiol. doi:10.1111/jam.13354 Kadkhoda, K., Manickam, K., DeGagne, P., Sokolowski, P., Pang, P., Kontzie, N., Alfa, M., 2011. UF-1000i flow cytometry is an effective screening method for urine specimens. Diagn. Microbiol. Infect. Dis. 69, 130–136. doi:10.1016/j.diagmicrobio.2010.09.013
PT
Le, Z., Li, F., Fei, C., Ye, A., Xie, X., Zhang, J., 2016. Performance of the Sysmex UF1000i urine analyser in the rapid diagnosis of urinary tract infections in
RI
hospitalized patients. J. Infect. Chemother. doi:10.1016/j.jiac.2016.02.009 Marschal, M., Wienke, M., Hoering, S., Autenrieth, I.B., Frick, J.S., 2012. Evaluation
Diagn. Microbiol. Infect. Dis. 72, 125–130.
NU
doi:10.1016/j.diagmicrobio.2011.10.001
SC
of 3 different rapid automated systems for diagnosis of urinary tract infections.
Medina-Bombardó, D., Jover-Palmer, A., 2011. Does clinical examination aid in the
MA
diagnosis of urinary tract infections in women? A systematic review and metaanalysis. BMC Fam. Pract. 12, 111.
Monsen, T., Rydén, P., 2015. Flow cytometry analysis using sysmex UF-1000i
D
classifies uropathogens based on bacterial, leukocyte, and erythrocyte counts in
PT E
urine specimens among patients with urinary tract infections. J. Clin. Microbiol. 53, 539–545. doi:10.1128/JCM.01974-14 Moshaver, B., de Boer, F., van Egmond-Kreileman, H., Kramer, E., Stegeman, C.,
CE
Groeneveld, P., 2016. Fast and accurate prediction of positive and negative urine cultures by flow cytometry. BMC Infect. Dis. 16, 211. doi:10.1186/s12879-016-
AC
1557-4
Muñoz-Algarra, M., Martínez-Ruiz, R., Orden-Martínez, B., 2013. Evaluación del sistema automatizado UF-1000i en el diagnóstico de infección urinaria. Enferm. Infecc. Microbiol. Clin. 31, 29–31. Pieretti, B., Brunati, P., Pini, B., Colzani, C., Congedo, P., Rocchi, M., Terramocci, R., 2010. Diagnosis of bacteriuria and leukocyturia by automated flow cytometry compared with urine culture. J. Clin. Microbiol. 48, 3990–3996. doi:10.1128/JCM.00975-10 Ronald, A., 2003. The etiology of urinary tract infection. Traditional and emerging pathogens. Dis Mon. 2003; 49:71–82. Schmiemann, G., Kniehl, E., Gebhardt, K., Matejczyk, M.M., Hummers-Pradier, E., 12
ACCEPTED MANUSCRIPT 2010. The diagnosis of urinary tract infection: a systematic review. Dtsch Arztebl Int 107, 361–367. doi:10.3238/arztebl.2010.0361 Shepherd, A.K., Pottinger, P.S., 2013. Management of Urinary Tract Infections in the Era of Increasing Antimicrobial Resistance. Med. Clin. North Am. 97, 737–757. doi:10.1016/j.mcna.2013.03.006 Shigemura, K., Arakawa, S., Tanaka, K., Fujisawa, M., 2009. Clinical investigation of isolated bacteria from urinary tracts of hospitalized patients and their
PT
susceptibilities to antibiotics. J. Infect. Chemother. 15, 18–22. doi:10.1007/s10156-
AC
CE
PT E
D
MA
NU
SC
RI
008-0652-x
13
ACCEPTED MANUSCRIPT Table 1. Demographic characteristics and urine culture data. n
N = 9057
% (95%CI)
Age (years)* Patients
47 (33.0 - 73.0) women
7726
85.3 (84.6 - 86.0)
men
1331
14.7 (14.0 – 15.4)
Primary health care
7034
77.7 (76.8 - 78.5)
Outpatients
1424
Emergency
368
Hospitalized patients
231
Culture results
2.6 (2.3 - 2.9)
41.1 (1 - 42.2)
2146
23.7 (22.8 - 24.6)
3186
35.2 (34.2 - 36.2)
Escherichia coli
1922
60.3 (58.6 – 62.0)
Klebsiella spp.
473
14.9 (13.7 - 16.1)
381
12.0 (10.9 - 13.3)
Proteus spp.
132
4.1 (3.5 - 4.9)
Other gram-negative bacilli
178
5.6 (4.8 - 6.4)
100
3.1 (2.6 - 3.8)
Contaminated
D
MA
Bacterial growth
Enterococcus spp.
CE
PT E
Urine pathogens
4.1 (3.7 - 4.5)
3725
NU
Negative
15.7 (15.0 - 16.5)
RI
SC
Urine sources
PT
Gender
Other microorganisms
AC
95% confidence interval of relative frequency (95%CI). *Median and interquartile range (IQR).
14
ACCEPTED MANUSCRIPT Table 2. Forming element counts in urine samples assessed with FCA organized by culture results. RBC (cells/µL)
EC (cells/µL)
E. coli (n=1922)
55.9 (0.2 - 33,243.2)
12.6 (0.1 - 6361.2)
4.6 (0.0 - 489.3)
Klebsiella spp. (n=473)
55.1 (0.4 - 32,777.6)
12.9 (0.5 - 1977.1)
4.2 (0.0 - 110.6)
Enterococcus spp. (n=381)
18.9 (0.2 - 14,869.8)
14,0 (0.9 - 515.0)
7.6 (0.0 - 157.0)
54.4 (1.0 - 1812.0)
11.3 (1.0 - 1962.9)
7.4 (0.0 - 129.0)
Other gram-negative bacilli (n=178) Other microorganisms (n=100)
43.6 (0.2 - 25,240.5)
RI
SC
Proteus spp. (n=132)
PT
WBC (cells/µL)
12.3 (0.5 - 755.2)
NU
N = 9057
4.4 (0.0 - 80.4)
20.8 (2.0 - 2106.7)
5.65 (0.1 - 259.2)
Contaminated urine (n=2146)
15.6 (0.0 - 14,234.1)
15.3 (0.2 - 37,118.2)
8.8 (0.0 - 519.0)
Negative urine (n=3725)
28.9 (0.1 - 32,621.6)
16.4 (0.2 - 37,593.2)
13.7 (0.0 - 962.7)
D
MA
70.9 (0.3 - 2799.7)
AC
CE
PT E
Data shown as median values and range. WBC, white blood cells; RBC, red blood cells; EC, epithelial cells
15
ACCEPTED MANUSCRIPT Table 3. Pair-wise comparison of urine culture results. Data shown as corrected pvalues (Benjamini-Hochberg method) with significance for p<0.05.
E. coli
Klebsiella
Enterococcus
Proteus
Other
spp
spp
spp
Gram-
E. coli
<0.001
Klebsiella spp
<0.001
0.371
Enterococcus spp
0.004
0.067
0.170
Proteus spp.
0.001
0.316
0.276
0.103
Other Gram-
<0.001
0.216
0.190
0.053
Others
0.030
<0.001
<0.001
0.002
Contamination
0.007
<0.001
0.002
0.095
Negative
E. coli
<0.001
<0.001
0.010
0.002
Klebsiella
Enterococcus
Proteus
Other
spp
spp
spp
Gram-
<0.001
Klebsiella spp
<0.001
0.239
Enterococcus spp
<0.001
<0.001
<0.001
Proteus spp.
<0.001
0.237
0.145
<0.001
Other Gram-
0.013
0.029
0.117
<0.001
0.028
Others
<0.001
0.343
0.235
<0.001
0.411
0.067
Contamination
<0.001
<0.001
<0.001
0.248
<0.001
<0.001
MA
NU
E. coli
Enterococcus
Proteus
Other
spp
spp
spp
Gram-
E. coli
E. coli
<0.001
D
Klebsiella
Negative
Klebsiella spp
<0.001
0.230
Enterococcus spp
<0.001
<0.001
<0.001
Proteus spp.
<0.001
<0.001
<0.001
0.469
Other Gram-
<0.001
0.310
0.499
<0.001
<0.001
<0.001
0.058
0.036
0.048
0.062
0.056
<0.001
<0.001
<0.001
0.091
0.251
<0.001
Others
<0.001 Others
0.004
AC
Contamination
CE
Others
PT E
EC
0.005
SC
WBC
Others
0.426
PT
Negative
RI
RBC
16
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Figure 1. Diagnostic accuracy with regard to pathogen growth in urine cultures according to bacteria and WBC count in flow cytometry. ROC curve analysis with different subsets of urine culture results. (A) All urine samples; (B) exclusion of contaminated urine samples; (C) only positive urine cultures with bacterial counts over 105 CFU/mL. Blue line (bacterial) and grey line (WBC).
17