Abstract
Objectives
Urinalysis is one of the most frequently ordered diagnostic laboratory tests. In order to reduce workload and costs, rapid screening tests such as urine test strip analyses are applied. The aim of this study was to evaluate the analytical performance of the UC‐3500 as well as the diagnostic performance in comparison with reference methods.
Design and methods
We measured within‐run and between‐run imprecision based on quantitative reflectance values. 347 prospectively included urine specimens were investigated for the presence of glucose, protein, albumin, leukocyte esterase, and hemoglobin peroxidase activity, and ordinal scale results were compared to an automated urine particle analyzer (UF‐5000, Sysmex, Kobe, Japan) and wet chemistry (Roche Cobas 8000, Mannheim, Germany).
Results
Within‐run and between‐run imprecision results based on reflectance data for both the 9 and 11 parameter test strips ranged from 0.07% to 1.36% for the low‐level control and from 0.37% to 6.13% for the high‐level control, depending on the parameter. Regarding diagnostic performance, the sensitivity/specificity for glucose, protein, albumin, leukocyte esterase, and hemoglobin peroxidase was 100/60%, 94.2/88.2%, 81.8/89.2%, 81.7/92.8%, and 85.1/88.6%, respectively; the negative predictive value was 100%, 83.3%, 89.1%, 94.6%, and 96.1%. The Spearman correlation coefficients of the UC‐3500 vs reference methods ranged from 0.915 to 0.967, depending on the parameter.
Conclusion
This fully automated urine test strip analyzer overall shows a satisfying performance and can reliably screen out negative urine samples in order to focus on further characterization of positive samples in the following steps of the workflow.
Keywords: dipstick, fully automated urine analyzer, test strip, urinalysis, urine chemistry analysis
Abbreviations
- CMOS
complementary metal oxide semiconductor detector
- CV
coefficient of variation
- FN
false negative
- FP
false positive
- Lc
confirmation limit
- LD
detection limit
- LG
Limit Grey zone
- NPV
negative predictive value
- PPV
positive predictive value
- R
reflectance
- r
Spearman’s rank correlation coefficient
- RBC
red blood cells
- RI
reference interval
- WBC
white blood cells
1. INTRODUCTION
Urinalysis is the third most requested analysis in clinical laboratories and can be used to identify patients with kidney diseases, urinary tract infections, or diabetes mellitus and can be part of their annual health checkup.1, 2 The high throughput of samples can create a significant workload, and a large proportion turns out to be negative.3 Therefore, it is of importance that results from early diagnostic examinations accurately rule out urine samples from further analysis by urine sediment analysis and culture.
The European Confederation of Laboratory Medicine has proposed a stepwise procedural strategy describing different levels.4 Level 1 methods are applied as first step, often delivering results on an ordinal scale. Urine dipstick analysis is used as a fast, first‐line screening method as part of a multistep workflow. As the reliability of visual inspection might be hampered by subjective color interpretation, instrumental reading is performed by automated analyzers.5
On most automated test strip readers, results are reported semiquantitatively. Due to the demand to analyze large sample volumes and in case a medical emergency situation implies the need for an instant urine status, semi‐automated or fully automated urine strip readers have been well accepted for standardized, high‐throughput screening. The obtained results should help to clearly separate samples without any indication for renal or genitourinary tract disorders from those samples with values exceeding the normal reference levels that need further examination. Positive samples might be subject for further microscopy, immunochemistry, or bacteriologic tests.
The recently introduced fully automated urine test strip reader UC‐3500 (Sysmex, Kobe, Japan) is designed to screen for 11 urinary parameters. Recently, it was demonstrated that the instrument shows an outstanding performance in screening for albuminuria.6 With an exceptionally low detection limit of 5.5 mg/L, this chemistry analyzer provides a very sensitive automated screening method. This is especially interesting as albumin levels between 20 and 200 mg/L act as an early indicator for vessel damage.7, 8, 9, 10 Also, the UC‐3500 reflectance data of leukocyte esterase and hemoglobin peroxidase showed good agreement with red blood cell (RBC) and white blood cell (WBC) counts obtained on the urine particle analyzer UF‐5000 (Sysmex, Kobe, Japan).11
In this study, we investigated the analytical and diagnostic performance of the UC‐3500 compared to reference methods. Imprecision measurements were based on quantitative reflectance data. The diagnostic performance was determined using the semiquantitative result categories implied by the manufacturer.
2. MATERIALS AND METHODS
2.1. Patient samples
Three hundred forty‐seven urine samples which were submitted to our laboratory between October 2016 and December 2016 were included. Routine diagnostic urinalysis and any additional study‐related procedures were performed on fresh urine specimens within 2‐4 hours after receipt.
2.2. Instrument and reagent strips
The fully automated urine test strip analyzer UC‐3500 (Sysmex, Kobe, Japan) was used for semiquantitative measurement of specified analytes in human urine and commercially available control materials according to the manufacturer's instruction.
Test strips (Meditape UC‐9A and UC‐11A, Sysmex, Kobe, Japan; Lot number: AC5004) were used in this study. These strips include reagent pads for ordinal scale reporting of urobilinogen, glucose, protein, hemoglobin peroxidase, nitrite, bilirubin, ketone, leukocyte esterase, pH, creatinine (UC‐11A), and albumin (UC‐11A) (Table 1). All steps starting from sample aspiration to reporting of results were obtained automatically. A maximum number of 300 urine test strips can be installed in the instrument at one time and up to 276 samples can be analyzed per hour. The instrument is equipped with a reflective photometry unit and reagent strips are scanned with a color complementary metal oxide semiconductor detector (CMOS), taking reflectance readings from the reagent strip. The light reflected off the reagent pad is used to measure the concentration of a substance present in the urine. A high concentration of analyte corresponds to a low reflectance. The reflectance value, expressed as a percentage within a range from 100% (white) to 0% (black), is inversely related to the concentration of the analyte. To cover the whole measuring range with good linearity, two overlapping reflectance ranges are installed for the parameters pH, protein, glucose, bilirubin, and ketones.
Table 1.
Parameter | Unit | Semiquantitative assessment categories | |||||
---|---|---|---|---|---|---|---|
Normal | ± | 1+ | 2+ | 3+ | 4+ | ||
Urobilinogen | mg/dL | 2.0 | 4.0 | 8.0 | 12.0 | ||
Glucose | mg/dL | ‐ | 50 | 100 | 250 | 500 | 2000 |
Protein | mg/dL | ‐ | 15 | 30 | 100 | 300 | 1000 |
Blood | |||||||
Red blood cells | cells/µL | ‐ | 10 | 20 | 50 | 250 | |
Hemoglobin | mg/dL | ‐ | 0.03 | 0.06 | 0.15 | 0.75 | |
Nitrite | ‐ | + | |||||
Bilirubin | mg/dL | ‐ | 0.5 | 1.0 | 2.0 | ||
Ketone | mg/dL | ‐ | 10 | 30 | 80 | ||
Leukocyte esterase | cells/µL | ‐ | 25 | 75 | 500 | ||
Creatinine | mg/dL | 10 | 50 | 100 | 200 | 300 | |
Albumin | mg/L | 10 | 30 | 80 | 150 | >150 | |
pH | NA | 5.0 5.5 | 6.0 6.5 | 7.0 7.5 | 8.0 8.5 | 9.0 |
2.3. Imprecision
Commercially available control material (UC‐control low [LOT: 01601‐L] and high [LOT: 01601‐H]; Sysmex, Kobe, Japan) was used to assess within‐run (n = 20) and between‐run (n = 20) imprecision on both UC‐9A and UC‐11A urine test strips. Intra‐run and between‐run imprecision were determined during one run on one day and on 20 consecutive days with one analysis a day, respectively.
2.4. Diagnostic accuracy
Sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) of the UC‐3500 vs defined reference methods were calculated for glucose (n = 59), protein (n = 67), hemoglobin peroxidase (n = 347), leukocyte esterase (n = 347), and albumin (n = 177) using laboratory‐specific reference intervals (RI) or predefined cutoffs.
2.5. Reference methods
Urinary glucose (RI: 0.0‐0.05 g/L), urinary creatinine (RI: 20‐400 mg/dL), and urinary total protein concentrations (cutoff 0.2 g/L) were determined on the Roche Cobas 8000 (Roche Diagnostics, Mannheim, Germany) using the hexokinase, Jaffe, and pyrogallol red‐molybdate (Instruchemie BV, Delfzijl, The Netherlands) photometric immunochemistry reference methods, respectively. As a comparison method for hemoglobin peroxidase and leukocyte esterase, RBCs and WBCs were counted on the Sysmex UF‐5000 fully automated urine particle analyzer (Sysmex Corporation, Kobe, Japan), respectively. The cutoff for both RBCs and WBCs was set at 25 cells/µL. The UF‐5000 showed very good agreement when compared to phase‐contrast microscopy using a Fuchs‐Rosenthal chamber.12 The Behring Nephelometer II analyzer (Siemens, Marburg, Germany) was used for assessment of urinary albumin concentration (cutoff 20 mg/L). The immunonephelometric assay was carried out using commercially available Siemens antibodies and the WHO/College of American Pathologists certified reference material 470.13, 14
2.6. Statistical analyses
All statistical analyses were performed by Excel 2013 (Microsoft, Seattle, WA, USA) and Analyse‐it ™ software, version 3.90.5 (Analyse‐it Software Ltd, Leeds, UK).
Semiquantitative results obtained by the UC‐3500 were compared to those of quantitative reference methods. The European Confederation of Laboratory Medicine provided guidance for the estimation of trueness of ordinal scale test strip evaluations when compared to other methods.2 Three analytical specifications zones and the allowance ranges of deviation were defined as follows: LD = detection limit below which a sample should be negative (false positives: optimum/minimum: <10/<20%); Lc = confirmation limit above which a sample should be positive (false negatives: optimum/minimum: <5/<10%); and LG = grey zone, between LD and LC (false negatives: optimum/minimum: <30/<50%). Agreement between reference methods and test strip data for glucose and protein was evaluated by Spearman rank regression analysis.
3. RESULTS
3.1. Imprecision
The within‐run imprecision ranged from 0.12% to 4.31% and 0.07% to 6.13% for the 9A and 11A urine test strip parameters, respectively. The between‐run imprecision ranged from 0.08% to 6.08% and 0.10 to 5.59% for the 9A and 11A urine test strip parameters, respectively. A summary of the imprecision results is presented in Table 2.
Table 2.
Parameter | UC‐control material | Meditape UC‐9A | Meditape UC‐11A | ||
---|---|---|---|---|---|
Within‐run CV, % | Between‐run CV, % | Within‐run CV, % | Between‐run CV, % | ||
Urobilinogen | Low | 0.12 | 0.17 | 0.20 | 0.20 |
High | 1.81 | 6.08 | 2.66 | 5.22 | |
Glucose | Low: R1/R2 | 0.56/0.56 | 0.65/0.62 | 0.45/0.45 | 0.60/0.57 |
High: R1/R2 | 4.31/2.24 | 4.80/3.07 | 3.18/2.11 | 4.55/2.73 | |
Protein | Low: R1/R2 | 0.23/0.77 | 0.29/1.00 | 0.29/0.71 | 0.23/0.74 |
High: R1/R2 | 0.81/1.20 | 1.29/3.00 | 0.86/1.90 | 1.04/2.27 | |
Blood | Low | 0.84 | 0.82 | 1.09 | 0.92 |
High | 2.25 | 3.31 | 3.58 | 4.24 | |
Nitrite | Low | 0.14 | 0.27 | 0.19 | 0.27 |
High | 1.17 | 1.04 | 0.98 | 1.04 | |
Bilirubin | Low: R1/R2 | 0.12/0.14 | 0.08/0.13 | 0.07/0.12 | 0.10/0.20 |
High: R1/R2 | 1.04/1.00 | 2.26/2.30 | 1.13/1.21 | 2.33/2.37 | |
Ketone | Low: R1/R2 | 0.10/1.04 | 0.16/1.36 | 0.17/1.15 | 0.11/1.04 |
High: R1/R2 | 0.92/3.10 | 1.03/5.69 | 0.98/6.13 | 0.95/5.59 | |
Leukocyte esterase | Low | 0.35 | 0.34 | 0.41 | 0.39 |
High | 2.11 | 4.00 | 2.09 | 3.27 | |
pH | Low: R1/R2 | 1.61/0.97 | 1.43/0.66 | 1.20/1.12 | 1.12/1.21 |
High: R1/R2 | 1.15/3.38 | 1.12/3.65 | 1.04/3.81 | 1.25/4.76 | |
Creatinine | Low | Not available | 0.71 | 0.93 | |
High | 3.97 | 5.13 | |||
Albumin | Low | Not available | 0.21 | 0.24 | |
High | 0.37 | 0.42 |
3.2. Comparison of biochemical reference method with test strip results
The correlation between urinary glucose concentrations and glucose oxidase on the test strip is presented in Figure 1. The following regression equations were obtained: y (1/reflectance (R1)) = 0.0433 + 0.0764 loggluC (g/L); Spearman r = 0.967; P < 0.001 (Figure 1A) and y (1/reflectance (R2)) = 0.0285 + 0.0500 loggluc (g/L); Spearman r = 0.915; P < 0.001 (Figure 1B).
The correlation between urinary protein concentrations and protein quantification on the urinary test strip is presented in Figure 2. The following regression equations were obtained for the UC‐9A test strips: y (1/reflectance (R1)) = 0.0099 + 0.0017 logprot (g/L); Spearman r = 0.939; P < 0.001 (Figure 2A) and y (1/reflectance (R2)) = 0.0138 + 0.0078 logprot (g/L); Spearman r = 0.944; P < 0.001 (Figure 2B) and the UC‐11A test strips: y (1/reflectance (R1)) = 0.0100 + 0.0017 logprot (g/L); Spearman r = 0.945; P < 0.001 (Figure 2C) and y (1/reflectance (R2)) = 0.0139 + 0.0076 logprot (g/L); Spearman r = 0.943; P < 0.001 (Figure 2D).
3.3. Diagnostic performance
The diagnostic accuracy of the UC‐3500 for urinary hemoglobin peroxidase, leukocyte esterase, glucose, protein, and albumin is presented in Table 3.
Table 3.
Parameter | Test strip | No. of Samples | Sensitivity (%) | Specificity (%) | NPVa (%) | PPVb (%) |
---|---|---|---|---|---|---|
Glucose | UC‐11A | 59 | 100 | 60 | 100 | 88 |
Protein | UC‐9A | 67 | 94.2 | 82.4 | 82.4 | 94.2 |
UC‐11A | 67 | 94.2 | 88.2 | 83.3 | 96.1 | |
Hemoglobin peroxidase | UC‐9A | 347 | 85.1 | 89.3 | 96.2 | 65.5 |
UC‐11A | 347 | 85.1 | 88.6 | 96.1 | 64.0 | |
Leukocyte esterase | UC‐9A | 347 | 80.5 | 91.7 | 93.8 | 75.0 |
UC‐11A | 347 | 81.7 | 92.8 | 94.6 | 77.9 | |
Albumin | UC‐11A | 177 | 81.8 | 89.2 | 89.1 | 81.8 |
Negative predictive value.
Positive predictive value.
An excellent NPV of 96.2% for RBCs and just below 95% for WBCs in comparison with flow cytometry was obtained. Both test strips provided similar results. For hemoglobin peroxidase and leukocyte esterase, we found a sensitivity of 85.1% (both test strips) and 80.5/81.7% (UC‐9A/UC‐11A), respectively (Table 3).
For glucose analysis, we determined a specificity of 60%. On the other hand, we found no negative results compared to the hexokinase method, yielding a sensitivity and NPV of 100%.
Regarding the determination of total protein, a sensitivity and specificity of 94.2% (both test strips) and 82.4/88.2% (UC‐9A/UC‐11A) were obtained, respectively.
For albumin measurement, a sensitivity of 81.8%, a specificity of 89.2%, a NPV of 89.1%, and a PPV of 81.8% were found for UC‐11A test strip analysis compared to the reference immunonephelometric method.
3.4. Agreement and evaluation of trueness
For the five parameters evaluated, an LD and LC were defined as follows: For RBC and WBCs, the LD and LC were set at 25 and 125 cells/µL, respectively; for glucose, the LD and LC were defined as 50 and 250 mg/dL, respectively; for albumin, the LD and LC were defined as 20 and 100 mg/dL, respectively; and for creatinine, the LD and LC were defined as 150 and 250 mg/dL. The agreement between the reference method and test strip analysis is presented in Table 4.
Table 4.
Parameter | Test strip | Perfect agreement (%), same category |
Agreement ±1 category (%) |
---|---|---|---|
Glucose | UC‐11A | 78.0 | 98.3 |
Protein | UC‐9A | 63.8 | 97.1 |
UC‐11A | 65.2 | 98.6 | |
Hemoglobin peroxidase | UC‐9A | 64.8 | 86.2 |
UC‐11A | 65.1 | 86.5 | |
Leukocyte esterase | UC‐9A | 77.2 | 96.5 |
UC‐11A | 78.1 | 97.1 | |
Albumin | UC‐11A | 79.1 | 100.0 |
Creatinine | UC‐11A | 53.9 | 95.7 |
The calculated FPD was well within the optimum criterion (<10%) for leukocyte esterase and within the minimum criterion (<20%) for the albumin, creatinine, and hemoglobin peroxidase. For protein and glucose, the FPD was not within the predefined criterion (Table 5).
Table 5.
Parameter | Test strip | FP at LD | FN in LG | FN at LC |
---|---|---|---|---|
Glucose | 50mg /dL | 50‐250mg /dL | >250mg /dL | |
UC‐11A | 40.0% (6/15) | 0.0% (0/20) | ND | |
Protein | <30mg /dL | 30‐150mg /dL | >150mg/dL | |
UC‐9A | 37.5% (6/16) | 44.7% (17/38) | 23.1% (3/13) | |
UC‐11A | 43.8% (7/16) | 44.7% (17/38) | 38.5% (5/13) | |
Hemoglobin peroxidase | <25 cells/µL | 25‐125 cells/µL | >125 cells/µL | |
UC‐9A | 10.7% (30/280) | 28.1% (9/32) | 2.9% (1/35) | |
UC‐11A | 11.4% (32/280) | 28.1% (9/32) | 2.9% (1/35) | |
Leukocyte esterase | <25 cells/µL | 25‐125 cells/µL | >125 cells/µL | |
UC‐9A | 8.3% (22/265) | 31.8% (14/44) | 5.3% (2/38) | |
UC‐11A | 7.2% (19/265) | 31.8% (14/44) | 2.6% (1/38) | |
Albumin | <20mg /dL | 20‐100mg /dL | >100mg /dL | |
UC‐11A | 10.0% (11/110) | 25.5% (12/47) | ND | |
Creatinine | <150mg /dL | 150‐250mg /dL | >250mg /dL | |
UC‐11A | 10.8% (10/93) | 8.3% (1/12) | 33.3 (1/3) |
In brackets, the absolute numbers of samples are displayed.
FP, false positive; FN, false negative, LD, detection limit; LG, limit grey zone; LC, confirmation limit; ND, not determined.
The calculated FNG was well within the optimum criterion (<30%) for albumin, creatinine, glucose, and hemoglobin peroxidase and within the minimum criterion (<50%) for protein and leukocyte esterase (Table 5).
Finally, the calculated FNC was well within the optimum criterion for hemoglobin peroxidase and leukocyte esterase (UC‐11A) and within the minimum criterion for leukocyte esterase (UC‐9A). For protein and creatinine, the FNC was not within the criterion (Table 5).
4. DISCUSSION
In the European Urinalysis Guidelines,2 four sequential diagnostic procedure levels with increasing accuracy were recommended for urinalysis. Level 1 methodologies represent fast screening tests often placed in primary care laboratories and at points‐of‐care. They should have a clinically acceptable performance to act as a sieving system in order to reduce the workload for higher levels. A good analytical and diagnostic accuracy in the lower range around the cutoff is especially important for primary screening technologies as they should reliably distinguish normal from positive samples. Only positive or suspicious samples would be subject for further investigation.
The application of test strips with multiple reagent pads belongs to the group of primary screening procedures. In 2016, the urine chemistry analyzer UC‐3500 was introduced in our laboratory. As the instrument is fully automated, interobserver variability associated with visual reading of test strips is overcome.
The within‐run and between‐run imprecision for the low‐level quality control, covering the negative and 1+ semiquantitative categories, were excellent. CVs for nine out of eleven parameters were below 1% and for ketones and pH just above 1%. A good analytical accuracy in the lower level range is especially important to distinguish normal or negative samples from positive samples. These results are comparable to other urinary test strip analyzers.15 CVs for within‐run and between‐run imprecision of high‐level control, covering 1+, 2+, 3+, and 4+ categories, ranged from below 1% (albumin) to 6% (ketones). This can be regarded as sufficient as the exact value in pathological samples will be determined by workflow level 2‐4 procedures.
The Spearman r‐values for glucose (0.96) and protein (0.90 and 0.92) were excellent and indicate a strong correlation with quantitative results. Earlier reports15 comparing hexokinase‐based glucose analyses and test strip reflectance readings on the URISYS 2400 (Roche Diagnostics, IN, USA) indicated a Spearman r of −0.85. In a recent publication, good correlation was found for flow cytometric WBC and UC‐3500 leukocyte esterase results (r = 0.82) as well as RBC and peroxidase (r = 0.84).11 Regression analysis of albumin measured on Meditape UC‐11A UC test strip versus immunonephelometry showed a strong correlation (r = 0.92), similar for creatinine (r = 0.90).6
Whereas the reference instrument UF‐5000 counts RBC and WBC via flow cytometry, the UC‐3500 measures the enzymatic activity of hemoglobin peroxidase and leukocyte esterase, respectively. It has been reported that due to low sensitivity and NPV, screening for urinary infections by test strips alone might not be sufficient.16 Despite these findings, in our study, the NPV was just below 95% for WBCs, and values for trueness of ordinal scale measurements fulfilled optimum criteria for the detection and confirmation limit. With 31.8% FN for the grey zone, the value was slightly above the optimum of 30%. In conclusion, the absence of urinary tract infections based on leukocytes can be reliably ruled out by test strip analysis using the UC‐3500. Results for the diagnostic accuracy for the detection of hematuria were also excellent: A NPV and sensitivity of 96.1% and 85.1% were obtained, respectively, and optimum criteria for evaluation of trueness were fully reached for Lc and LG; for LD, the value was slightly above the optimum 10% FP.
Screening for glycosuria by urine dipstick analysis may identify patients with undetected diabetes mellitus. Sensitivity and NPV of UC‐3500 vs Cobas 8000 glucose measurement were 100%, and there were no false‐negative cases above the detection limit. Below the LD, 40% FP were found and the minimum criterion of <50% FP was met. Detection of undiagnosed diabetes by urinary glucose screening happens only by chance and is therefore of minor relevance. Both the detection of diabetes and monitoring of a respective therapy are superior in blood testing.
Proteinuria is defined as the excretion of more than 150 mg of protein per day, a hallmark of renal disease, and an indicator for hypertension in pregnant women. Repeatability for total protein was acceptable with ≤1% CV for low‐level and ≤3% CV for high‐level control material, along with an excellent sensitivity (94.2%) and PPV (94.2/96.1). The NPV was well above 80% for the two test strips. In a previous publication, a correlation coefficient of 0.97 for a range of 15‐1000 mg/dL was described, reflecting a nearly perfect positive relationship between UC‐3500 reflectance data and Cobas 8000 immunochemistry results for proteinuria.6
Regarding albumin measurement, all optimum criteria for trueness were fulfilled and the performance for ordinal scale measurements confirmed the ones recently published for nominal scale measurements.9
In conclusion, the fully automated test strip analyzer UC‐3500 provides a high‐throughput first‐level screening method for urinalysis which acts as a reliable sieving system to reduce the workload for further validation methods.
ACKNOWLEDGMENTS
We thank Dr. Elisabeth Breit, PhD, for her valuable contribution to and review of the article.
Oyaert M, Delanghe JR. Semiquantitative, fully automated urine test strip analysis. J Clin Lab Anal. 2019;33:e22870 10.1002/jcla.22870
REFERENCES
- 1. Coppens A, Speeckaert M, Delanghe J. The pre‐analytical challenges of routine urinalysis. Acta Clin Belg. 2010;65:182‐189. [DOI] [PubMed] [Google Scholar]
- 2. European Urinalysis Guidelines . Scand. J Clin Lab Invest Suppl. 2000;231:1‐86. [PubMed] [Google Scholar]
- 3. Oyaert M, Delanghe J. Progress in automated urinalysis. Ann Lab Med. 2019;39:15‐22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Kouri TT, Gant VA, Fogazzi GB, Hofmann W, Hallander HO, Guder WG. Towards European urinalysis guidelines. Introduction of a project under European Confederation of Laboratory Medicine. Clin Chim Acta. 2000;297:305‐311. [DOI] [PubMed] [Google Scholar]
- 5. Rowell DM. Evaluation of a urine chemistry analyser. Prof Nurse. 1998;13:533‐534. [PubMed] [Google Scholar]
- 6. Delanghe JR, Himpe J, De Cock N, et al. Sensitive albuminuria analysis using dye‐binding based test strips. Clin Chim Acta. 2017;471:107‐112. [DOI] [PubMed] [Google Scholar]
- 7. Svendsen PA, Oxenboll B, Christiansen JS. Microalbuminuria in diabetic patients: a longitudinal study. Acta Endocrinol Suppl (Copenh). 1981;242:53‐54. [PubMed] [Google Scholar]
- 8. Levey AS, Eckardt KU, Ysukamoto Y, et al. Definition and classification of chronic kidney disease: a position statement from kidney disease: improving global outcomes (KDIGO). Kidney Int. 2005;67:2089‐2100. [DOI] [PubMed] [Google Scholar]
- 9. Eknoyan G, Hostetter T, Bakris GL, et al. Proteinuria and other markers of chronic kidney disease: a position statement of the National Kidney Foundation (NKF) and the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). Am J Kidney Dis. 2003;42:617‐622. [DOI] [PubMed] [Google Scholar]
- 10. Smink PA, Lambers HJ, Gansevoort RT, et al. Albuminuria, estimated GFR, traditional risk factors, and incident cardiovascular disease: the PREVEND (Prevention of Rena land Vascular ENdstage Disease) study. Am J Kidney Dis. 2012;60:804‐811. [DOI] [PubMed] [Google Scholar]
- 11. Oyaert M, Himpe J, Speeckaert M, Stove V, Delanghe J. Quantitative urine test strip reading for leukocyte esterase and hemoglobin peroxidase. Clin Chem Lab Med. 2018;56:1126‐1132. [DOI] [PubMed] [Google Scholar]
- 12. Previtali G, Ravasio GR, Seghezzi M, Buoro S, Alessio MG. Performance evaluation of the new fully automated urine particle analyser UF‐5000 compared to the reference method of the Fuchs‐Rosenthal chamber. Clin Chim Acta. 2017;472:123‐130. [DOI] [PubMed] [Google Scholar]
- 13. Fink PC, Römer M, Haeckel R, et al. Measurement of proteins with the behring nephelometer. A multicenter evaluation. J Clin Chem Clin Biochem. 1989;27:261‐276. [PubMed] [Google Scholar]
- 14. Decavele AS, Fiers T, Penders J, Delanghe JR. A sensitive quantitative test strip based point of care albuminuria assay. Clin Chem Lab Med. 2012;50:673‐678. [DOI] [PubMed] [Google Scholar]
- 15. Penders J, Fiers T, Delanghe JR. Quantitative evaluation of urinalysis test strips. Clin Chem. 2002;48:2236‐2241. [PubMed] [Google Scholar]
- 16. İnce FD, Ellidağ HY, Koseoğlu M, Şimşek N, Yalçın H, Zengin MO. The comparison of automated urine analyzers with manual microscopic examination for urinalysis automated urine analyzers and manual urinalysis. Pract Lab Med. 2016;11:14‐20. [DOI] [PMC free article] [PubMed] [Google Scholar]