You are on page 1of 6

Microbiology and Infectious Disease / UF-1000i for UTI Diagnosis

Evaluation of the Sysmex UF-1000i for the Diagnosis of Urinary Tract Infection
Jie Wang,1 Ying Zhang,2 DongWen Xu,1 WeiJun Shao,1 and Yuan Lu1
Key Words: Sysmex UF-1000i; Flow cytometry; Urinary sediment examination; Urinary tract infection
DOI: 10.1309/AJCP1GT2JXOCQBCZ

Abstract
Diagnosis of urinary tract infection (UTI) is primarily done by microbiologic culture, which is timeconsuming and can produce false-positives and falsenegatives. Flow cytometry allows for rapid screening of many samples and eliminates culturing. We analyzed the Sysmex UF-1000i (TOA Medical Electronics, Kobe, Japan) for accuracy in identifying RBCs and WBCs, casts, bacteria, and epithelia. We also evaluated its precision, linear estimation of results, carryover contamination rate, and anti-interference. UF-1000i agreement with manual counting was approximately 95% for RBCs and WBCs, epithelia, and casts. Its coefficient of variation for bacteria ranged from 4.7% to 15.2%. UF-1000i screening for UTIs exhibited great sensitivity (97%), specificity (79%), positive predictive value (70%), negative predictive value (99%), and accuracy (85%). The negative predictive value remained high even with complex UTI samples. The Sysmex UF-1000i shows great promise in excluding more than 50% of true-negative samples, improving detection efficiency, and reducing laboratory costs.

Urinary tract infection (UTI) has been cited as the most common bacterial infection, and it may have serious complications for children, people with diabetes, elderly people, and people with immune compromise.1,2 It is diagnosed primarily by urine culture and clinical manifestations, with microbiologic culture still the diagnostic gold standard,3 despite being time-consuming, labor-intensive, and prone to false-negative or false-positive results.4-6 Flow cytometry has long been recognized as capable of identifying bacteria,7 and the Sysmex UF-100 automated urine particle flow cytometer (TOA Medical Electronics, Kobe, Japan), in use since 1997, has shown increasing sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy in screening for UTI.4,8-11 Recently, the UF-100 was updated with an improved model, Sysmex UF-1000i. The UF-1000i uses a semiconductor laser ( 635 nm) with forward and side scatter detection and a single DNA dye for its 2 counting channels, 1 for sediment analysis and 1 for microbes. When the fluorescein-stained urine sample is inserted into the UF-1000i sheath-flow cross-flow pool, it is irradiated by the laser beam, and particles in the urine sample produce signals in the forward scattering light, side scattering light, and side fluorescence signal. Those signals are converted to opticoelectronic signals so components can be identified, counted, classified, and analyzed. Bacteria are counted via a special bacterial examination channel, and interference with RBCs is prevented. We hypothesized that, compared with its predecessor, the UF-1000i would have better sensitivity, specificity, PPV, NPV, and accuracy in identifying the 5 formed components of urine (RBCs, WBCs, casts, bacteria, and epithelial cells) and
Am J Clin Pathol 2010;133:577-582 577
DOI: 10.1309/AJCP1GT2JXOCQBCZ

American Society for Clinical Pathology

577 577

Wang et al / UF-1000i for UTI Diagnosis

evaluated its precision, linear estimation of results, carryover contamination rate, and anti-interference.

Materials and Methods


Specimens Random urine samples (n = 313) were obtained from inpatients and outpatients of our hospital from August 8, 2007, to March 17, 2008. During the same period, the microbiology room of the hospital Department of Laboratory Medicine obtained 368 urine samples from various departments for urine culture, using disposable urine collection cups with covers. Sample collection was performed in strict accordance with the operational standards of clean midstream urine collection.3 The study was approved by the hospital institutional review board and was in accordance with the Helsinki Declaration of 1975. No informed consent was required. Urine Culture for UTI Diagnosis All samples were inoculated within 1 hour after collection. Clean midstream urine was fully mixed and was inoculated onto the blood plate and MacConkey flat plate using a 0.001-mL quantitative inoculation ring. Samples were cultured in an incubator at 37C for 24 hours, at which time colony counting and Gram staining were carried out. Gram-negative bacteria of more than 105 colony-forming units (CFU)/mL and gram-positive bacteria of more than 104 CFU/mL indicated a UTI, as per domestic12 and National Committee for Clinical Laboratory Science13 standards for diagnosis. This is more rigorous than the international standard of more than 105 CFU/mL for gram-negative and gram-positive bacteria.14 Manual Examination of Urine Samples Two senior technicians used the FAST-READ 10 counting board (Immune Systems, London, England) to count RBCs, WBCs, casts, bacteria, and epithelial cells manually in accordance with the urine sample examination standard.15 The RBC and WBC counts were divided into 7 levels: 0 to 50, 51 to 100, 101 to 200, 201 to 300, 301 to 500, 501 to 800, and 801 to 1,000/L. The epithelial cell count was divided into 4 levels: fewer than 15, 15 to 25, 29 to 49, and 50 to 100/ L. The cast count was divided into 4 levels: 0 to 3, 3.1 to 6, 6.1 to 9, and 9.1 to 12/L. Technicians were blinded to the UF-1000i results. Examination of Urine RBCs, WBCs, Casts, Bacteria, and Epithelial Cells Using the UF-1000i For analysis, a 10-mL urine sample was collected and put into a test tube. The test tube was put into the matching test-tube rack of the instrument, and the instrument examined
578 578 Am J Clin Pathol 2010;133:577-582
DOI: 10.1309/AJCP1GT2JXOCQBCZ

the samples automatically. Cells were differentiated using multiple indexes, including the forward scattering light signal, the side scattering light signal, and fluorescence signal, and cells of similar size, such as crystals and RBCs and also small round cells and WBCs were identified and classified. The pathologic casts and casts without inclusions are differentiated according to the level of fluorescence distribution width. The machine provides an automatic count of RBCs, WBCs, casts, bacteria, and epithelial cells. UF-1000i for UTI Diagnosis To determine the amount of gram-negative bacteria and gram-positive bacteria, we converted the domestic standard for a diagnosis of UTI to one appropriate for the UF-1000i. To do this, we first set the colony count of the quantitative bacterial culture of the regular clean midstream urine sample (with urine that has stayed in the bladder for 4-6 hours) at 105/mL or more, set the bacterial count at 105/mL or more, and set the yeast-like fungi count at 100/L or more. Then we set the number of WBCs in the sediment of the centrifuged clean midstream urine at more than 10/HP for samples from patients with UTI symptoms. Because the UF-1000i analyzer can convert between HP and microliters (1:5.6), the number of WBCs was set to more than 56/L. Examination of UF-1000i Precision The RBCs, WBCs, casts, bacteria, and epithelial cells in the 313 random fresh urine samples were examined manually, then examined using UF-1000i for the aforementioned 5 components. Each sample was continuously examined 11 times. The results of the last 10 examinations were recorded and analyzed. Linear Estimation of the UF-1000i Examination Results A high-value urine sample with values close to the expected upper limit (RBCs, 10,967/L; WBCs, 5,100/L) was selected and diluted at ratios of 1:4, 1:16, 1:64, 1:256, and 1:1,024 in negative urine (all indexes near 0). The measured values were compared with the theoretical values, and the correlation coefficient was used to estimate linear correlation. Examination of the Carryover Contamination Rate of the UF-1000i One fresh urine sample with high values and one fresh urine sample with low values were randomly selected and the aforementioned 5 indexes determined by UF-1000i. In the high-value urine sample, components were as follows: RBCs, 10,807/L; WBCs, 3,197/L; casts, 23/L; bacteria, 91,301/ L; and epithelial cells, 516/L. The low-value fresh urine sample contained the following: RBCs, 27.7/L, WBCs, 38.0/ L, casts, 1/L, bacteria, 92/L, and epithelial cells, 2/L. To assay carryover contamination, first, the high-value sample
American Society for Clinical Pathology

Microbiology and Infectious Disease / Original Article

was analyzed 3 times continuously using the UF-1000i. Then, the low-value sample was examined 3 times without modifying any instrument settings or changing the pipette. The carryover contamination rate was evaluated by calculating the effect of the high-value sample examination on the results of the low-value sample, using the following formula: wx yz 100%

Table 1 Values of RBC, WBC, Epithelial Cell, and Cast Counts Using the UF-1000i and the Quantitative Counting Plate Method
Agreement (%) RBCs WBCs Epithelial cells Casts 95.11 96.70 94.15 95.79

0.83 0.91 0.81 0.26

where w equals the first-time result of the low-value sample, x equals the third-time result of the low-value sample, y equals the third-time result of the high-value sample, and z equals the first-time result of the high-value sample. Evaluation of the Anti-interference Performance of the UF-1000i Venous RBCs were added into a normal fresh urine sample to prepare the hematuria sample with maximum expected concentrations of RBCs. The actual concentration of RBCs was counted manually using a counting board as described. This hematuria sample was diluted in normal urine using the doubling dilution method (dilution rate of 2). Interfering factors (bacteria and yeast-like fungi) were added to the samples at different dilution rates, and the UF-1000i was used to measure the RBC count. The actual measured RBC count was compared with expected values to evaluate the antijamming performance of the UF-1000i. Statistical Analysis The Cohen was used on dichotomous data to establish that agreement exceeded that expected under the null hypothesis of random ratings. Yerushalmy models were constructed to indicate presence of bacteria as causative for UTI.16 The correlation coefficient was used to estimate the linear correlation of theoretical vs actual counts of RBCs and WBCs as measured by the UF-1000i. Data were analyzed using STATA version 9.0 (StataCorp, College Station, TX), and a P value less than .05 indicated statistical significance.

a value of 0.26, suggesting little correlation between the 2 methods Table 1. UF-1000i for Identifying Gram-Positive and GramNegative Bacteria Among the 368 urine culture specimens, 127 showed positive results, for a positive rate of 34.5%. Among the 132 bacterial strains isolated and identified by microbial culture, 66 were gram-negative bacteria (49.3%), 36 were gram-positive bacteria (26.9%), and 30 were fungi (22.4%). The most common microorganisms identified by the UF-1000i from a urine sample (in order of frequency) were as follows: Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, and Proteus mirabilis (gram-negative bacteria); Enterococcus faecium, Enterococcus faecalis, Staphylococcus haemolyticus, Staphylococcus epidermidis, and Staphylococcus aureus subsp aureus (gram-positive bacteria); and Candida albicans and Candida tropicalis (yeast). UF-1000i UTI detection was compared with the results of microbial culture. The sensitivity, specificity, PPV, NPV, and accuracy of UF-1000i for UTI diagnosis were calculated (WBC >56/ L or bacteria >105/mL or yeast-like fungi >100/L).7 The sensitivity was 0.86, specificity was 0.95, PPV was 0.91, NPV was 0.94, and the consistency rate was 0.91. UF-1000i Precision Precision of the UF-1000i in identifying the 5 formed components in urine is provided in Table 2. The coefficient of variation (CV) for detecting RBCs was 3.5% to 5.3% (25.16-206.80/L); for detecting WBCs, it was 1.7% to 8.9% (20.93-488.30/L); for bacteria, it was 4.7% to 15.2% (110.40-78,781.00/L); for casts, it was 8.4% to 16.0% (5.64-21.31/L); and for epithelial cells, it was 4.8% to 19.7% (13.88-119.33/L). Linear Assessment of UF-1000i Examination The correlation coefficient was used to estimate linear correlation of UF-1000i Figure 1. The linear estimate of the UF-1000i for RBCs was r = 1.000 (P < .05), and for WBCs, it was r = 1.000 (P < .01).
Am J Clin Pathol 2010;133:577-582 579
DOI: 10.1309/AJCP1GT2JXOCQBCZ

Results
UF-1000i vs Manual Counting All clinical urine samples (n = 313) examined by UF-1000i and manual counting were considered consistent if the difference between the results of the 2 examination methods was within 1 SD. The RBC consistency rate was 95.11%, the WBC consistency rate was 96.70%, and the epithelial cell consistency rate was 94.15%. Except for casts, all values were more than 0.81, indicating good correlation between the 2 methods. The consistency rate for casts was 95.79% with
American Society for Clinical Pathology

579 579

Wang et al / UF-1000i for UTI Diagnosis

Table 2 SD, Mean, and Coefficient of Variation of Cells


Cell Concentration (/L) RBCs <30 <80 <300 WBCs <30 <80 <300 <500 Bacteria 105-106 1-2 106 <107 Epithelial cells <20 <120 SD Mean Coefficient of Variation (%)

interfering factors of 100 to 400/L did not cause statistically significant interference with RBC examination (P < .0001).

1.33 3.17 7.15 1.87 2.31 9.18 8.30 16.73 66.49 4,590.98 2.73 5.68

25.16 72.92 206.80 20.93 78.43 272.10 488.30 110.40 1,399.80 78,781.00 13.88 119.33

5.3 4.3 3.5 8.9 2.9 3.4 1.7 15.2 4.7 5.8 19.7 4.8

Discussion
Urinary sediment examination is an important component of urine analysis, greatly assisting with clinical diagnosis, therapeutic monitoring, and health screenings. In our analysis, the UF-1000i achieved a CV for RBCs of at least 5.3%. This correlates well with the recent analysis by Shayanfar et al,17 who found rates of 6.5% and 5.5% for the Iris iQ200 (Iris Diagnostics, Chatsworth, CA) and the UF-100, respectively. We similarly found the newer flow cytometer to outperform the earlier device. Ozdem et al5 found CVs of 6.35% to 12.18% for WBCs using the UF-100 at different concentrations; our range was 1.7% to 8.9%. In our study, sensitivity was 86.0%, specificity was 95.0%, PPV was 91.0%, NPV was 94.0%, and consistency was 91%. False-negatives can result from other systemic infections, antibiotic use close to the time of the test, and delay in processing.4-6 For the dipstick test, false-negatives can result from abnormal pH levels, proteinuria, the presence of certain metabolites or vitamin C, and even test strip error.18,19 The cultures of the 2 false-negative samples were S aureus and E faecium. The UF-1000i bacterial counts were 104 to 105/mL. Although those numbers met the clinical criteria of positive infection with cocci, they did not reach the positive value set in this study and were considered negative. Among the 52 false-positive samples, the WBC count exceeded 56/L in 22 samples, perhaps because microbial culture of the urine

The Carryover Contamination Rate of the UF-1000i The carryover contamination rates of RBCs, WBCs, epithelial cells, casts, and bacteria were 0.07%, 0.03%, 0.02%, 0.49%, and 0.04%, respectively. Anti-interference Performance of UF-1000i Examination Results of the hematuria sample examined by UF-1000i with bacterial interference are shown in Table 3. Correlation coefficient analysis showed no difference among theoretical values in RBC at interference levels of 104 to 107/mL (P < .0001). Results of the hematuria sample examined by the UF-1000i with fungal interference are shown in Table 4. Correlation coefficient analysis showed that concentrations of

A
12,000 RBC Theoretical Value/L

B
6,000 WBC Theoretical Value/L 5,000 4,000 3,000 2,000 1,000 0 0 2,000 4,000 6,000 8,000 10,000 12,000 RBC Measured Value/L 0 1,000 2,000 3,000 4,000 5,000 6,000 WBC Measured Value/L

10,000 8,000 6,000 4,000 2,000 0

Figure 1 Theoretical vs actual RBC (A) and WBC (B) counts. The theoretical values were determined by the correlation coefficient and the actual count from UF-1000i readings.

580 580

Am J Clin Pathol 2010;133:577-582


DOI: 10.1309/AJCP1GT2JXOCQBCZ

American Society for Clinical Pathology

Microbiology and Infectious Disease / Original Article

Table 3 Bacterial Interference With the RBC Count*


Concentration of Interfering Bacteria
Expected Concentration of RBCs (/L) 1 104/mL 1 105/mL 1 106/mL 1 107/mL

Table 4 Interference of Yeast-like Fungi With the RBC Count*


Concentration of Yeast-like Fungi (L) Expected Concentration of RBCs (/L) 50 1,082 540.8 210.7 100.2 54.2 24.9
*

100 1,106.7 535.8 226.3 131.9 71 46.3

200 1,119.7 573.9 247.1 142.5 96.4 68.3

400 1,158.9 603.9 297.8 186 129 101.1

940 467.2 192.1 91.8 39.2 19.5


*

988.9 467.5 198.7 103.4 46.8 20.2

979.3 466.8 184.9 96.8 48.7 23.4

988.2 469.1 191.5 95.9 47.2 23.7

974.5 470.4 185.9 97.7 49.2 20.4

1,101.4 521.6 212.8 113.1 66 34.6

Results are given as means.

Results are given as means.

failed, the patient had been taking antibiotics before sampling, or there was a lack of bacteriuria, sometimes reported in UTI.2,4 In addition, the instrument gave a review signal for 10 samples, suggesting a certain degree of interference with the analysis. In a study of the UF-100 to analyze peritoneal dialysis fluid, Penders et al20 combined WBC count, bacterial channel count, and total protein measurement to find a sensitivity of 75.0% (range, 47.6%-92.6%) and a specificity of 72.2% (range, 46.5%-90.2%) for diagnosing peritonitis. As others have suggested,4,16 the value of the UF-1000i lies not in replacing laboratory analysis, but in reducing the number of samples that require laborious manual counting. The UF-1000i shows great diagnostic potential for UTI screening,14 diagnosing other renal and urologic conditions,21 and examination of even cerebrospinal22 and other fluids. Urinary sediment analysis requires a variety of equipment and methods, and interfering factors can produce great analytic error and differences between laboratories and slow automation progress. Three national urine analysis seminars resulted in a proposal for standardized urinary sediment analysis and equipment.23 According to the report of the urine analyzer manual developed by the Clinical and Laboratory Standards Institute, European Confederation of Laboratory Medicine,13,24 and Japanese Research Institute on Urinary Tract Infection, results of urinary sediment examination should be reported quantitatively. The UF-1000i automated urine particle flow cytometer can meet these requirements and classify a large number of urine particles quickly.25 In this study, the results of the UF-1000i performance evaluation showed that this instrument has good precision, accuracy, carryover contaminationprevention ability, and linear and antijamming performance, also consistent with microscopic counting. In this sample, UF-1000i identified 65.5% of 386 samples as negative, in keeping with standards seen in the literature for the UF-100.4,6,25,26 The consistency rate for casts was 95.79% with a of 0.26, suggesting little correlation between the
American Society for Clinical Pathology

two methods. This result may have been influenced by the high proportion of samples with low cast values and the low proportion with high cast values, which hampered statistical understanding (Table 2). The sensitivity of the UF-1000i for UTI screening was 97%, the specificity was 79%, the PPV was 70%, the NPV was 99%, and the accuracy was 85%. The NPV was high when the results of WBC, bacteria, and yeast-like fungi were combined for UTI screening, which may help the clinical laboratory filter out more than 50% of true-negative samples, improving detection efficiency and reducing laboratory costs.
From the 1Department of Laboratory Medicine, Huashan Hospital, Fudan University; and 2Sysmex Shanghai Scientific and Application Department, Shanghai, China. Address reprint requests to Dr Lu: Dept of Laboratory Medicine, Huashan Hospital, Fudan University, 12 Wulumuqi Zhong Rd, Shanghai 200040, China. Acknowledgment: We are grateful to Longquan Gu, MD, Ruijin Hospital, Shanghai Jiaotong University, and Daming Jin, MD, Shanghai Clinical Laboratory Center for assistance in the evaluation of the UF-1000i; and we thank the microbiology room staff of our Department of Laboratory Medicine for their support.

References
1. Foxman B. Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Am J Med. 2002;113:5-13. 2. Juthani-Mehta M, Tinetti M, Perrelli E, et al. Diagnostic accuracy of criteria for urinary tract infection in a cohort of nursing home residents. J Am Geriatr Soc. 2007;55:1072-1077. 3. Collins CH, Lyne PM. Microbiological Methods. 5th ed. London, England: Butterworths; 1985:128-134. 4. Manoni F, Valverde S, Antico F. Field evaluation of a second-generation cytometer UF-100 in diagnosis of acute urinary tract infections in adult patients. Clin Microbiol Infect. 2002;10:662-668.

Am J Clin Pathol 2010;133:577-582 581


DOI: 10.1309/AJCP1GT2JXOCQBCZ

581 581

Wang et al / UF-1000i for UTI Diagnosis

5. Ozdem S, Bayraktar T, Oktay C, et al. The prevalence of asymptomatic pyuria in diabetic patients: comparison of the Sysmex UF-100 automated urinalysis analyzer with FuchsRosenthal hemacytometer. Clin Biochem. 2006;39:873-878. 6. Zaman Z, Roggeman S, Verhaegen J. Unsatisfactory performance of flow cytometer UF-100 and urine strips in predicting outcome of urine cultures. J Clin Microbiol. 2001;39:4169-4171. 7. Van Dilla MA, Langlois RG, Pinkel D, et al. Bacterial characterization by flow cytometry. Science. 1983;220:620-622. 8. DeKeijzer MH, Brandts RW. Flow cytometry and the urine laboratory: field evaluation of the Sysmex UF-100. Sysmex J Int. 1997;7:1-7. 9. Fenili D, Pirovano B. The automation of sediment urinalysis using a new urine flow cytometer (UF-100). Clin Chem Lab Med. 1998;36:909-917. 10. Kouri TT, Khknen U, Malminiemi K, et al. Evaluation of Sysmex UF-100 urine flow cytometer vs chamber counting of supravitally stained specimens and conventional bacterial cultures. Am J Clin Pathol. 1999;112:25-35. 11. Lou Z, Hu X, Jiang Y. Clinical significance of the use of UF-100 urine analyzer for the urinary tract infection screening. Lab Med. 2004;19:54-56. 12. Ye Y, Wang M, Shen Z. National Clinical Examination Procedures. 3rd ed. Beijing, China: Medical Secretary, Ministry of Health, Peoples Republic of China; 2006:744. 13. National Committee for Clinical Laboratory Standards. Urinalysis and Collection, Transportation, and Preservation of Urine Specimens; Approved guideline. Villanova, PA: NCCLS; 1995:51-69. NCCLS Document GP-16A. 14. Grosso S, Bruschetta G, De Rosa R, et al. Improving the efficiency and efficacy of pre-analytical and analytical workflow of urine cultures with urinary flow cytometry. New Microbiol. 2008;31:501-505. 15. Itokoki I. Activity report of the committee of urine examination standardization. J Jap Clin Exam Stand Assn. 2005;20:18-20.

16. Yerushalmy J, Palmer CE. On the methodology of investigations of etiologic factors in chronic diseases. J Chronic Dis. 1959;108:27-40. 17. Shayanfar N, Tobler U, von Eckardstein A, et al. Automated urinalysis: first experiences and a comparison between the Iris iQ200 urine microscopy system, the Sysmex UF-100 flow cytometer and manual microscopic particle counting. Clin Chem Lab Med. 2007;45:1251-1256. 18. Simerville JA, Maxted WC, Pahira JJ. Urinalysis: a comprehensive review. Am Fam Phys. 2005;71:1153-1162. 19. Winstock GB. Manufacturing fault causes a false negative for haematuria using Makromed m81 urinalysis test strips: a case report. Med Sci Law. 2009;49:56-59. 20. Penders J, Fiers T, Dhondt AM, et al. Automated flow cytometry analysis of peritoneal dialysis fluid. Nephrol Dial Transplant. 2004;19:463-468. 21. Penders J, Fiers T, Everaert K, et al. Diagnostic performance of combined specific urinary proteins and urinary flow cytometry in urinary tract pathology. Clin Chem Lab Med. 2007;45:499-504. 22. Nanos NE, Delanghe JR. Evaluation of Sysmex UF-1000i for use in cerebrospinal fluid analysis. Clin Chim Acta. 2008;392:30-33. 23. Cong Y. Proposal for the urine sediment examination standardization. Chin J Lab Med. 2002;25:249-250. 24. ECLM (European Confederation of Laboratory Medicine) European Urinalysis Group. European urinalysis guidelines. Scand J Clin Lab Invest. 2000;20:231. 25. Hannemann-Pohl K, Kampf SC. Automation of urine sediment examination: a comparison of the Sysmex UF-100 automated flow cytometer with routine manual diagnosis. Clin Chem Lab. 1999;37:753-764. 26. Gessoni G, Valverde S, Maturi P, et al. Cytometry in the diagnosis of acute urinary tract infections: one years experience. Scand J Clin Lab Invest. 2004;64:77-80.

582 582

Am J Clin Pathol 2010;133:577-582


DOI: 10.1309/AJCP1GT2JXOCQBCZ

American Society for Clinical Pathology

You might also like