Clinical Chemistry / REDUCING URINE MICROSCOPY ANALYSES

Safely Reducing Manual Urine Microscopy Analyses

by Combining Urine Flow Cytometer and Strip Results
Sylvie Roggeman, MD, and Zahur Zaman, MD, PhD

Key Words: Flow cytometer; UF-100; Urine microscopy; Urine strips

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

Abstract Microscopic examination and chemical analysis of
We aimed to reduce the number of manual urine urine remain essential for diagnosis, prognosis, and follow-
microscopy examinations safely by cross-interpretation up of patients with renal and urinary tract diseases.
of the Sysmex UF-100 (TOA Medical Electronics, Kobe, However, urine sediment microscopy, being time
Japan) and urine strip results such that microscopy consuming and labor intensive, adds considerably to the
would be performed if there was discordance between cost of providing laboratory services. Attempts have been
the UF-100 and urine strip results. We also evaluated made to reduce the number of urine microscopic analyses
the usefulness of the optional UF-100 expert software. by screening the urine samples with urine strips and
We performed 2 studies: study 1 to establish review performing microscopy on the samples that give positive
rules for eventual microscopic examination; study 2, a results. Such simplification has been reported to produce
validation study. Our review rates were 40% and 48% significant losses in diagnostic yields.1-3 Another approach
and those of UF-100 software were 16% and 32% for to increase productivity has been to automate urine
the 2 studies. Our false-positive and false-negative microscopy on an image analysis system 4 and, more
results, among the samples not flagged for microscopic recently, on a Sysmex UF-100 flow cytometer (TOA
review, were acceptably low. We did not find a good Medical Electronics, Kobe, Japan). The latter counts the
correlation between the microscopic classification of formed elements in urine on the basis of their light scat-
RBC morphologic features and the classification given tering, fluorescence, and impedance properties. The princi-
by the UF-100. Since incorporation of the automated ples of analysis5 and evaluation6-9 have been published. The
urine strip reader and the UF-100 in routine use, our analyzer counts RBCs, WBCs, squamous epithelial cells,
manual microscopy has been reduced to less than 40%. bacteria, and casts but reports the presence of pathologic
casts, tubular and transitional epithelial cells (small round
cells), yeast-like cells, crystals, and spermatozoa semiquan-
titatively as positive. Now, an optional expert system has
been added to the software. This expert system interprets
the urine strip and UF-100 results and then generates
suggestive interpretative messages (flags), which may be
used as review rules for manual microscopy. The results of
the unflagged specimens would not require verification by
manual microscopy and could be validated and released,
should one so choose.
We performed a study to examine the possibility of
safely reducing the number of manual microscopy analyses

872 Am J Clin Pathol 2001;116:872-878 © American Society of Clinical Pathologists

 Clinical Chemistry / ORIGINAL ARTICLE

by cross-interpretation of the results of the UF-100 flow Menarini Diagnostics, Florence, Italy) for microscopic
cytometer and of an automated urine strip reader. We evalu- counting and examination of the formed elements. This was
ated the diagnostic accuracy of the UF-100 and its optional followed by strip analysis on an automated urine strip
expert software by comparing all UF-100 and the strip analyzer, Super Aution, using Uriflet S 9UB strips (A.
reader results with those of the reference method, manual Menarini Diagnostics). This strip enables semiquantitative
microscopy. This enabled us to generate our own review analyses of glucose, protein, bilirubin, urobilinogen, pH,
rules, determine what would be the diagnostic loss if heme (hemoglobin/myoglobin), nitrite, and leukocyte
manual microscopy were not performed on the specimens esterase. Super Aution measures specific gravity by the
not flagged for microscopic review, and establish whether refractive index method, color by reflectance measurement at

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

any UF-100 results would be released that were inconclu- 4 wavelengths, and turbidity by a transparency index
sive or apparently uninterpretable because UF-100 results method. It uses the Sysmex racks as sample holders. After
were positive and the urine strip reader results were nega- Super Aution the sample racks were transferred to the
tive or vice versa and, thus, mutually discordant for certain Sysmex UF-100 urine flow cytometer for quantitative
parameters. analysis of RBCs, WBCs, squamous epithelial cells,
bacteria, and casts and semiquantitation of pathologic casts,
tubular and transitional epithelial cells (small round cells),
yeast-like cells, crystals, and spermatozoa.
Materials and Methods
Urine Microscopy
Patients and Samples Manual microscopy was used as the “gold standard.”
Microscopic identification and counting were performed by
Study 1 phase-contrast microscopy (×400) on uncentrifuged urine
During a 4-week period, 1,029 freshly collected urine samples using disposable Uriglass counting chambers. Each
samples of hospital inpatients submitted for urine 1-µL chamber is divided into 10 large squares. Each square
microscopy were randomly selected for study 1. The patients has a volume of 0.1 µL and is subdivided into 16 small
consisted of 470 females (mean age, 50 years; range, 1 squares. RBCs, WBCs, cylinders, and tubular cells were
month to 96 years) and 559 males (mean age, 47 years; counted in 5 × 16 squares. The results for RBCs and WBCs
range, 1 month to 87 years). The specimens originated from were expressed as the number of cells per microliter of urine
nephrology (n = 321; of these, 132 were renal transplant (our reference values for RBCs, 6/µL or less; for WBCs,
recipients); pediatrics (n = 161); surgery (n = 36); intensive <10/µL) and those for casts and tubular cells as trace, 1+, or
and emergency departments (n = 105); oncology (n = 24); 2+. Other cells, crystals, and bacteria were counted in 2 × 16
internal medicine, including consultation (n = 278); geri- squares, and the results also were expressed as trace, 1+, and
atrics (n = 43); and other clinical departments (n = 61). 2+. For each type of formed element, the semiquantitative
scale corresponds to a certain number per microliter.
Study 2
The review and release rules generated in study 1 were Sysmex UF-100
validated in study 2 using 554 other urine samples. The We determined the microscopic review rate by using
patients consisted of 284 females (mean age 52 years; range, the review messages (flags) generated by the UF-100
1 month to 95 years) and 270 males (mean age, 56 years; expert software. The flags are as follows: (1) “high total
range, 1 week to 93 years). The source distribution was count” when the total number of particles is 250,000/µL or
similar to that of study 1. more; (2) conductivity of the sample is too high or too low;
(3) problem with morphologic discrimination of RBCs; (4)
Urine Collection and Analysis “myoglobin?/lysed RBC?” when hemoglobin/myoglobin
Random urine, voided in a plastic beaker, was aspi- detected by the strip is disproportionately more than the
rated (10 mL) into Uridraw evacuated conical plastic tubes RBCs found by the UF-100; (5) “hematuria?” when the
(Terumo Europe, Leuven, Belgium) and sent to the labora- number of RBCs per microliter is disproportionately much
tory by means of a pneumatic tube system (Aerocom greater than the hemoglobin value; (6) “pathologic cylin-
GmbH & Co, Kernen, Germany). For very young infants, ders?” when pathologic cylinders are more than 1/µL; (7)
urine was first collected in a plastic bag attached to the “casts?” if hyaline cylinders are more than 3/µL; (8) “old
external genital area. sample?” if bacteria per microliter are proportionately
After mixing, an aliquot of the urine sample was applied much more than WBCs per microliter; (9) “sterile pyuria?”
to a disposable plastic Uriglass counting chamber (A. when the WBCs per microliter are proportionately much

© American Society of Clinical Pathologists Am J Clin Pathol 2001;116:872-878 873

Roggeman and Zaman / REDUCING URINE MICROSCOPY ANALYSES

❚Table 1❚
Our Rules for Generating Flags That Led to Microscopy of the Flagged Specimen*

Formed Element UF-100 Result Strip Result

RBCs >30/µL and Hemoglobin, <0.06 µg/dL

<20/µL and Hemoglobin, ≥0.06 µg/dL
≥20/µL and ≤30/µL and Hemoglobin, <0.03 µg/dL or >0.1 µg/dL
WBCs >30/µL and <25/µL
<20/µL and ≥25/µL
≥20/µL and ≤30/µL and <25/µL or >75/µL
Hyaline cylinders >3/µL and Protein, <100 mg/dL

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

Pathologic cylinders ≥1/µL —
Small round cells >2/µL —
Squamous epithelial cells >50/µL —
Yeast-like cells >30/µL —
High total count >250,000/µL —
Conductivity ≤5 millisiemens/cm or ≥38 millisiemens/cm —
* Reference limits in the UF-100 software (Sysmex, TOA Medical Electronics, Kobe, Japan) were as follows: RBC, <25/µL; WBC, <25/µL; hyaline cylinders, <3/µL; pathologic
cylinders, <1/µL; yeast, <10/µL; small round cells, <1/µL; and epithelial cells, <30/µL.

more than bacteria per microliter; and (10) “proteinuria?” Results

when protein detected by the strip is more than 120 mg/dL.
It was assumed that in routine use of the UF-100 expert Comparing UF-100 Expert Software and Our Own
software, the results of the unflagged specimens (whether Review Rules
positive or negative) would be deemed not to require
microscopic review and thus would be validated and With the review rules of Sysmex UF-100 expert soft-
released. The reference limits in the UF-100 software were ware and with our own rules generated in study 1, 16.1%
as follows: RBCs, fewer than 25/µL; WBCs, fewer than (166/1,029) and 39.7% (408/1,029) of all samples, respec-
25/µL; hyaline cylinders, fewer than 3/µL; pathologic cylin- tively, would have required manual microscopy. In the
ders, fewer than 1/µL; yeast cells, fewer than 10/µL; small validation study (study 2), the review rate with our review
round cells, fewer than 1/µL; and squamous epithelial cells, rules was 48% and for Sysmex UF-100 expert software it
fewer than 30/µL. was 32%.
We also generated our own review rules ❚Table 1❚, based
on the analysis of manual microscopy, UF-100, and urine Red Blood Cells
strip reader data. If there was concordance between the strip An ROC curve was constructed for the UF-100 data on
and UF-100 results, microscopy was deemed unnecessary, RBC counts using manual microscopy as the reference
and the results were validated and released. When there was method. The area under the curve was 0.83 (95% confidence
a discordance between the UF-100 and strip results, the interval, 0.81-0.86). On this curve, an RBC cutoff of 20/µL
sample underwent microscopy. We defined discordance as on the UF-100 corresponded to a sensitivity of 58% and a
complete disagreement between the 2 results, ie, one is posi- specificity of 89%.
tive and the other is negative. In addition, when the numbers Manual microscopy was positive for RBCs (ie, > 6/µL)
of pathologic casts (casts with inclusions) and yeast-like in 443 of 1,029 cases. For the same 1,029 samples, the
cells were greater than our cutoffs of 1/µL or more and 50/µL Sysmex expert software generated 83 (8.1%) “hematuria?”
or more, respectively, the sample also was examined under a review flags. If microscopy were not done on the unflagged
microscope. One sample could be flagged for more than 1 samples, as would be the case in routine use, 186 (18.1%)
parameter. would be reported as false-negative results ❚Table 2❚. In 135
cases (13.1%), inconclusive results would be released since
Statistical Analysis in these cases there was mutual discordance between the UF-
Diagnostic sensitivities and specificities, receiver oper- 100 and the strip results (ie, one was positive and the other
ating characteristic (ROC) curves, correlations, and regres- was negative).
sion calculations were performed with the aid of the With our own review rules, 35% of 1,029 samples were
Analyse-It computer package (Analyse-It Software, Leeds, flagged for manual microscopy. Of the 621 unflagged
England). In all calculations, manual microscopy was used samples (60.3%), 565 were concordant by all 3 methods,
as the reference method. with 102 being positive and 463 negative ❚Table 3❚. There

874 Am J Clin Pathol 2001;116:872-878 © American Society of Clinical Pathologists

 Clinical Chemistry / ORIGINAL ARTICLE

❚Table 2❚
Diagnostic Accuracy, Relative to Manual Microscopy, of Results for Samples Not Flagged by UF-100 Expert Software*

Study 1 Study 2

Parameter TP (%) TN (%) FN (%) FP (%) TP (%) TN (%) FN (%) FP (%)

RBCs 116 (11.3) 515 (50.0) 186 (18.1) 16 (1.6) 87 (15.7) 207 (37.4) 61 (11.0) 23 (4.2)
WBCs 157 (15.3) 564 (54.8) 56 (5.4) 56 (5.4) 93 (16.8) 236 (42.6) 22 (4.0) 27 (4.9)
Hyaline cylinders 1 (0.1) 778 (75.6) 46 (4.5) 8 (0.8) 1 (0.2) 333 (60.1) 42 (7.6) 2 (0.4)
Pathologic cylinders 0 (0) 823 (80.0) 7 (0.7) 3 (0.3) 0 (0) 369 (66.6) 9 (1.6) 0 (0)
Yeast 0 (0) 821 (79.8) 11 (1.1) 1 (0.1) 0 (0) 375 (67.7) 3 (0.5) 0 (0)

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

Small round cells 37 (3.6) 328 (31.9) 10 (1.0) 458 (44.5) 22 (4.0) 130 (23.5) 3 (0.5) 223 (40.3)
Squamous epithelial cells 18 (1.7) 749 (72.8) 15 (1.5) 51 (5.0) 7 (1.3) 334 (60.3) 2 (0.4) 35 (6.3)

FN, false negative; FP, false positive; TN, true negative; TP , true positive.
* For study 1, n = 833 specimens not flagged of 1,029 total specimens; for study 2, n = 378 specimens not flagged of 554 total specimens. Percentages are based on the total

numbers of specimens. In routine use, these results would be released without microscopic review. For proprietary information, see the footnote for Table 1.

❚Table 3❚
Diagnostic Accuracy, Relative to Manual Microscopy, of Results for Samples Not Flagged by Our Review Rules*

UF-100 Results

Study 1 Study 2

Parameter TP (%) TN (%) FN (%) FP (%) TP (%) TN (%) FN (%) FP (%)

RBCs 102 (9.9) 463 (45.0) 53 (5.2) 3 (0.3) 85 (15.3) 174 (31.4) 20 (3.6) 5 (0.9)
WBCs 138 (13.4) 462 (44.9) 12 (1.2) 9 (0.9) 86 (15.5) 185 (33.4) 6 (1.1) 7 (1.3)
Hyaline cylinders 0 (0) 617 (60.0) 4 (0.4) 0 (0) 0 (0) 283 (51.1) 1 (0.2) 0 (0)
Pathologic cylinders 0 (0) 620 (60.3) 1 (0.1) 0 (0) 1 (0.2) 283 (51.1) 0 (0) 0 (0)
Yeast 1 (0.1) 619 (60.2) 1 (0.1) 0 (0) 0 (0) 281 (50.7) 3 (0.5) 0 (0)
Small round cells 11 (1.1) 553 (53.7) 3 (0.3) 54† (5.2) 13 (2.3) 229 (41.3) 1 (0.2) 41† (7.4)
Squamous epithelial cells 9 (0.9) 597 (58.0) 8 (0.8) 7 (0.7) 1 (0.2) 272 (49.1) 2 (0.4) 9 (1.6)

FN, false negative; FP, false positive; TN, true negative; TP , true positive.
* For study 1, n = 621 specimens not flagged of 1,029 total specimens; for study 2, n = 284 specimens not flagged of 554 total specimens. Percentages are based on the total

numbers of specimens. In routine use, these results would be released without microscopic review. For proprietary information, see the footnote for Table 1.
† When the UF-100 reference value for small round cells is raised to 2/µL, 54 is reduced to 26 and 41 to 22.

were 3 false-positive results (0.3% of the total) and 53 false- own review rules, 217 samples (21.1%) required manual
negative results (5.2% of the total). Of the latter, 3 were clin- microscopy for WBCs. Of the 621 (60.3% of the total) not
ically significant. But no inconclusive results were reported. requiring manual microscopy for any parameter, 603 were
In the validation study (study 2), among the 284 unflagged concordant with microscopy results; 138 were positive and
samples, there were 5 false-positive results (0.9% of the 462 negative (Table 3). There were 12 false-negative results,
total) and 20 false-negative results (3.6% of the total) for and of these, 4 were clinically significant. In study 2, 213
RBCs (Table 3); only 2 (0.4% of total) of the latter were clin- (38.7%) of 554 samples were positive for WBCs. Among
ically significant. the unflagged samples, there were 7 false-positive and 6
false-negative results (Table 3); only 2 of the latter were
White Blood Cells clinically significant.
The ROC curve for the UF-100 data for WBCs, with
manual microscopy as the reference method, yielded an area Casts
under the curve of 0.88 (95% confidence interval, 0.86-0.91). In study 1, manual microscopy was positive for hyaline
From this curve, a WBC cutoff of 20/µL on the UF-100 gave cylinders in 56 (5.4%) of 1,029 cases. The UF-100 expert
diagnostic sensitivity and specificity of 82% and 83%, software flagged 72 samples (7.0% of the total) for
respectively. microscopy; 48 flags were for “casts?” and 24 for “protein?”
By manual microscopy, 321 (31.2%) of 1,029 study 1 Of the 833 samples (81% of the total) not flagged by the
samples were positive for WBCs. The Sysmex UF-100 UF-100 expert software for any of the parameters, 539 gave
expert system does not have review rules, other than the concordant results by microscopy, UF-100, and protein; all
reference limit of fewer than 25/µL, for WBCs. Using our were negative. Eight were false-positive results, 46 were

© American Society of Clinical Pathologists Am J Clin Pathol 2001;116:872-878 875

Roggeman and Zaman / REDUCING URINE MICROSCOPY ANALYSES

false-negative results (Table 2), and in 270 cases (26.2% of 120

the total), inconclusive results would have been reported

Distribution Width (Channels)

100
because the UF-100 and the strip results were discordant.
With our review rules, among the unflagged samples, 80
there were 4 false-negative results for hyaline cylinders, no
false-positive results, and no inconclusive results (Table 3). 60

For the pathologic cylinders, 18 (1.8% of the total) were 40

positive by manual microscopy. Among the unflagged
samples, our review rules yielded 1 false-negative result and 20

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

no false-positive results in study 1 (Table 3). By comparison,
0
the UF-100 software produced 7 false-negative and 3 false- 0 20 40 60 80 100 120 140 160 180
positive results (Table 2). Forward Scatter (Channels)
In study 2, manual microscopy yielded only 6 positive
❚Figure 1❚ Scattergram of RBC morphologic features as
cases for pathologic cylinders. Among the unflagged
determined by the UF-100 (for proprietary information, see the
samples, with our rules, there were no false-negative or false-
footnote for Table 1) and microscopic examination. UF-100
positive results (Table 3), but with the UF-100 software
defines forward scatter (Fsc) >100 and distribution width (DW)
rules, there were 9 false-negative and no false-positive results
<50 as “isomorphic” (area enclosed by thick lines); Fsc <80 as
(Table 2).
“dysmorphic” (area to the left of the broken thick line), and the
Epithelial Cells rest as “mixed.” Microscopically determined morphologic
features (dysmorphic RBCs) are represented as follows:
By microscopic examination, 63 (6.1%) of 1,029
closed diamond, 0%-9%; closed square, 10%-19%; closed
samples were positive for tubular/bladder epithelial cells
triangle, 20%-29%; x, 30%-39%; open square, 40%-49%;
(small round cells) in study 1, as were 48 (8.7%) of 554 in
closed circle, 50%-59%; +, 60%-69%; open triangle, 70%-
study 2. With a cutoff of 1 small round cell on the UF-100,
79%; open circle, 80%-89%; and open diamond, 90%-100%.
the 2 studies yielded 458 and 223 false-positive results,
respectively, among the unflagged samples (Table 2). These
could be safely reduced to 54 or 26 for study 1 and 41 or 21
for study 2 by raising the cutoff to 2 small round cells per values. In 214 samples, with 25 RBCs per microliter or more
microliter, respectively (Table 3). by the UF-100, the percentages of isomorphic and dysmor-
For squamous epithelial cells, 44 (4.3%) of 1,029 phic RBCs were determined under the microscope, and the
samples were positive by microscopy in study 1, as were 13 results were compared with those of the UF-100 ❚Figure 1❚.
(2.3%) of 554 in study 2. With a cutoff of 30 squamous There was no acceptable relation between the microscopi-
epithelial cells per microliter, among the unflagged samples, cally determined percentage of dysmorphic RBCs and their
there were 51 false-positive results in study 1 and 35 in study UF-100 classification.
2 (Table 2). These false-positive results were reduced to 7
and 9, respectively, for studies 1 and 2 when the cutoff was
raised to 50/µL (Table 3).
Discussion
Yeast We evaluated the feasibility of safely reducing the
By manual microscopy, 39 (3.8%) of 1,029 samples number of manual microscopic analyses by cross-checking
were positive in study 1, as were 25 (4.5%) of 554 in study the urine strip and UF-100 results such that concordant
2. Among the samples not flagged for review by the UF-100 results would be released but samples showing discordance
software, there were 11 false-negative results in study 1 and between the strip and UF-100 results or being flagged for
3 false-negative results in study 2 (Table 2). Our review infringement of a review rule would be reviewed manually
rules for yeasts gave 1 false-negative result and no false- under a microscope. In this way, only conclusive results
positive results among the nonreview samples (Table 3). In would be released. To make the process acceptable to local
study 2, 3 false-negative results were found among the clinicians, our review rules were deliberately made quite
unflagged samples. stringent. With our rules, the review rate in study 1 was 40%
(of all the microscopy requests) and in study 2 was 48%.
RBC Morphologic Features With the review rules of the UF-100 software, the review
The UF-100 detects morphologic features by the size of rates for the respective studies were 16% and 32%. Our
the RBC as defined by forward scatter and distribution width review rates also were higher than those of Lun et al.10 The

876 Am J Clin Pathol 2001;116:872-878 © American Society of Clinical Pathologists

 Clinical Chemistry / ORIGINAL ARTICLE

explanation for our high review rate lies in the extensive- Since incorporation of the automated urine strip reader
ness of our review rules and limited scope of the UF-100 and the UF-100 in our urine workstation, manual microscopy
software and of the review rules of Lun et al.10 The Sysmex has been reduced to less than 40%. A substantial number of
UF-100 software does not flag all discrepancies between these are requests for analysis of RBC morphologic features.
the strips and the UF-100, and, other than the reference This reduction in manual workload has enabled us to reduce
limit, it contains no review rules for WBC counts. This daily occupancy of the urine workstation by 2 full-time
means that not only is the review rate lower, but also many equivalent technologist positions.
discrepant and therefore inconclusive results are released.
For their review criteria Lun et al 10 looked only at the From the Department of Laboratory Medicine, University

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

discrepancies in RBC and WBC counts between the UF- Hospital Leuven, Leuven, Belgium.
100 and the urine strips. Address reprint requests to Dr Zaman: Dept of Laboratory
The review rates for the Sysmex and by our own rules in Medicine, University Hospital Leuven, Herestraat 49, B-3000
study 2 were much greater than those in study 1. One Leuven, Belgium.
possible explanation for this may be that in study 2, there Acknowledgments: We thank Merck Eurolab (Leuven,
were more samples from kidney and pancreas transplant Belgium) and A. Menarini Diagnostics (Zaventem, Belgium) for
providing the apparatus and the reagents.
recipients and patients undergoing chemotherapy. Zaman et
al11 observed, for example, that, because of the presence of
pancreatic esterases, urine samples from kidney and pancreas
transplant recipients give discordant results for leukocyte References
esterase and WBC counts.
1. Valenstein PN, Koepke JA. Unnecessary microscopy in
With microscopy as the gold standard, our false-nega- routine urinalysis. Am J Clin Pathol. 1984;82:444-448.
tive rates in both studies were acceptably low for all the para- 2. Shaw ST Jr, Poon SY, Wong ET. “Routine urinalysis”: is the
meters except perhaps for RBCs (Table 2). Closer examina- dipstick enough? JAMA. 1985;253:1596-1600.
tion of the cases revealed that, except for 3 cases in study 1 3. Resnick M. Comment on simplifying urinalysis. Clin Chem.
and 2 cases in study 2, we would not have missed anything 1985;31:450-451.
clinically significant. The high false-positive rates for the 4. Roe CE, Carlson DA, Daignault RW, et al. Evaluation of
the yellow IRIS: an automated method for urinalysis. Am J
small round cells (epithelial cells) and the squamous epithe- Clin Pathol. 1986;86:661-665.
lial cells (Table 2) were due to low UF-100 reference limits 5. Nakomoto H. Automated urinalysis. Sysmex J Int.
for these parameters (1/µL for small round cells and 30/µL 1996;6:168-172.
for squamous epithelial cells). These values can safely be 6. Fenili D, Pirovano B. The automation of sediment
raised to 2 or 3/µL for small round cells and to 50/µL for urinalysis using a new urine flow cytometer (UF-100). Clin
squamous epithelial cells. When this is done, the false-posi- Chem Lab Med. 1998;36:909-917.
tive rates for these parameters drop, respectively, to 26 and 7 7. Ben-Ezra J, Bork L, McPherson RA. Evaluation of the
Sysmex UF-100 automated urinalysis analyzer. Clin Chem.
for study 1 and 22 and 9 for study 2 (Table 3). 1998;44:92-95.
The UF-100 also provides morphologic differentiation 8. Langlois MR, Delange JR, Steyaert SR, et al. Automated
of the RBC as isomorphic, dysmorphic, or mixed popula- flow cytometer compared with automated dipstick reader
tion. This feature is intended as an aid to diagnosing for urinalysis. Clin Chem. 1999;45:118-122.
glomerular hematuria12 and is based on the commonly held 9. Kouri TT, Kahkonen U, Malmaniemi K, et al. Evaluation of
Sysmex UF-100 urine flow cytometer vs chamber counting
belief that dysmorphic RBCs are glomerular in origin.13,14 of supravitally stained specimens and conventional
This idea is no longer wholly sustainable since dysmorphic bacterial cultures. Am J Clin Pathol. 1999;112:25-35.
cells occur in urine samples of patients without glomeru- 10. Lun A, Ziebig R, Priem F, et al. Routine workflow for use of
lonephritis15,16 and even in people with normal renal func- urine strips and urine flow cytometer in the hospital
laboratory. Clin Chem. 1999;45:1305-1307.
tion.17 The acanthocytes or G1 cells described as markers
11. Zaman Z, Roggeman S, Cappelletti P, et al. Evaluation of
of glomerular hematuria18-20 have good diagnostic speci- Aution Max AX-4280 automated urine test-strip analyser.
ficity but poor sensitivity.15,21 In the present study, unlike Clin Chem Lab Med. 2001;39:649-657.
the findings of Lun et al,10 we found no clearly discernible 12. Hyodo T, Kumano K, Haga M, et al. Detection of
relation between the UF-100 classification and the micro- glomerular and non-glomerular red blood cells by
scopically determined percentages of isomorphic and automated urinary sediment analyzer. Jpn J Nephrol.
1995;37:35-43.
dysmorphic RBCs (Figure 1). At this stage, we do not
13. Birch DF, Fairley KF. Haematuria: glomerular or non-
make use of the algorithm for RBC morphologic features glomerular [letter]? Lancet. 1979;2:845-846.
of the UF-100. Instead, we perform morphologic analysis 14. Fairley KF, Birch DF. Hematuria: a simple method for
under a microscope. identifying glomerular bleeding. Kidney Int. 1982;21:105-108.

© American Society of Clinical Pathologists Am J Clin Pathol 2001;116:872-878 877

Roggeman and Zaman / REDUCING URINE MICROSCOPY ANALYSES

15. Zaman Z, Proesmans W. Dysmorphic erythrocytes and G1 19. Tomita M, Kitamoto Y, Nakayama M, et al. A new
cells as markers of glomerular hematuria. Pediatr Nephrol. morphological classification of urinary erythrocytes for
2000;14:980-984. differential diagnosis of glomerular hematuria. Clin Nephrol.
16. Pollock C, Pei-Ling L, György AZ, et al. Dysmorphism of 1992;37:84-89.
urinary red blood cells: value in diagnosis. Kidney Int. 20. Lettgen B, Wohlmuth A. Validity of G1-cells in the
1989;36:1045-1049. differentiation between glomerular and nonglomerular
17. Hyodo T, Kumano K, Haga M, et al. Analysis of urinary red haematuria in children. Pediatr Nephrol. 1995;9:435-437.
blood cells of healthy individuals by automated flow 21. Gyory AZ, Hawkins T, Ross M, et al. Clinical value of urine
cytometer. Nephron. 1997;75:451-457. microscopy by manual and automated methods. Lab Hematol.
18. Kohler H, Wandel E, Brunck B. Acanthocyturia: a 1998;4:211-216.
characteristic marker for glomerular bleeding. Kidney Int.

Downloaded from https://academic.oup.com/ajcp/article-abstract/116/6/872/1758291 by Joongbu University user on 11 April 2020

1991;40:115-120.

878 Am J Clin Pathol 2001;116:872-878 © American Society of Clinical Pathologists