The standardization of the report for urine cell counting—A converting factor for Sysmex UF‐1000i

Background Multicenter laboratory may apply both automated flow cytometer and microscopy for urinalysis. Automated flow cytometer such as Sysmex UF‐1000i evaluates particles with native urine without centrifugation and reports as “counts per μL.” Microscopic examination recommended as the reference method for urine sediment analysis reports results as “counts per HPF (or μL).” Moreover, some results from flow cytometer are needed to be checked visually under microscopy. Therefore, it is worth to establish the consistency of the results from these two methods. Methods Urine specimens from 412 patients were examined with Sysmex UF‐1000i and manual microscopy using FAST‐READ disposable counting chambers. White blood cell (WBC) and red blood cell (RBC) counting results from UF‐1000i after transferred with the converting factor (0.297) we estimated were compared with that from microscopic examination. Method comparison was performed using Passing‐Bablok analysis. Results After transferred with the converting factor (0.297), cell counting results from UF‐1000i showed a good correlation with that derived by the reference method (R 2 was 0.868 for RBCs (P < 0.001), 0.882 for WBCs (P < 0.001)). Passing‐Bablok analysis showed no systematic difference (intercept estimate, −1 [95%CI, −7 to 3] and slightly proportional (slope estimate, 1.2 [95%CI, 1.0 to 1.7]) bias between concentrations of cells measured by manual microscopy and Sysmex UF‐1000i using the converting factor. Conclusion The converting factor (0.297) helps to transfer “counts per μL (non‐centrifugal urine)” to “counts per μL (equal to centrifugal urine),” and to keep the urine particle analysis results of Sysmex UF‐1000i consistent with the results from the reference method.

brings confusion to clinicians when they are not used to the units of UF series. 3 The microscopic urine sediment analysis is recommended as the reference method for the urine cell analysis. 4,5 The sample showing abnormalities or no coincident between the flow cytometer and the dipstick was re-examined under the microscopy using cell chambers (KOVA or FAST-READ etc). The guideline GP-163A 4

approved by Clinical and Laboratory Standards
Institute (CLSI) mentioned that the urinalysis results should be reported in the same reporting format and using the same reference intervals, and emphasized that the cell counts from manual microscopy are not suitable for comparison between laboratories. However, Chinese expert consensus 5 required that results from different methods-based analyzers should be reported with their own reference interval and proposed that the urine sediment results should be reported as "counts per HPF (centrifugal urine)." In summary, the agreement has not been reached of these guidelines on the unit for reporting urine sediment results. The various reporting forms for the same test may bring confusion to clinicians who evaluate the treatment effect for patients with kidney disease on the basis of urine sediment results.
We adopted a converting factor to make the results from UF-1000i comparable to the results from recommended urine sediment examined procedure. After that, we testified this converting factor by an experiment.

| Specimens
A total of 412 fresh urine specimens were obtained from inpatients of Guangdong Provincial Hospital of Chinese Medicine during a period of 3 weeks. The samples were collected in polypropylene containers. The analysis was performed on the UF-1000i (Sysmex Corporation, Kobe, Japan) and then by FAST-READ disposable counting chambers (Immune Systems Ltd, Paignton, UK). This whole process was finished within 1 hour of receipt.

| Centrifugation efficiency estimation
The percentage of particles remained in supernatant or lost by discard could only be estimated. Fifty-five urine specimens were randomly selected for centrifugation efficiency estimation. After analyzed on the UF-1000i for the first time, samples were centrifuged for 5 minutes at 400 g. The supernatant was absorbed into a new tube (leaving about 200 μL residue) and mixed by reversing the tubes for eight times and then measured by UF-1000i again.
The centrifugation efficiency was calculated by the equation below:

| Adjustment of the UF-1000i converting factor
According to the recommended procedure of urine sediment analysis, 4 the sediment samples were examined without staining, and the RBC and WBC enumeration results were reported as average particle counts per high-power field (HPF, ie, 1 field at 400× magnification, 10× 0bjective).

| Urine particle analysis with Sysmex UF-1000i analyzer and microscopy
The urine samples was centrifuged at 400 g for 5 minutes, then the

| Statistical analysis
The SPSS statistical 20.0 software (IBM, Armonk, NY, USA) was used to analyze the EF results and to do regression analysis between the results from UF-1000i and chamber methods. EF data were presented as X50% (X25%, X75%). Passing-Bablok analysis was also performed to make comparison between UF-1000i and chamber methods (after conversion).

| Efficiency of centrifugation
The efficiency of centrifugation is presented in Figure 1. Around the concentration of the reference limit, the efficiency factor of RBC (EFR) was 77.8% (10%, 99%), and the efficiency factor of WBC (EFW) was 88.9% (29%, 100%). The EF was relatively high at the high concentration of cells. The centrifuge efficiency should be carefully considered.  Table 1.

| Comparison results between FAST-READ chambers and UF-1000i after conversion
The 95% confidence range for the intercept of the regression lines contains 0, which means there is no systematic difference between UF-1000i (using the converting factor) and chamber counting. The slopes of the estimated linear functions, at around 1.2, are very close to 1.0, indicating a good correlation between two methods.

| D ISCUSS I ON
Sysmex UF-1000i (TOA Medical Electronics, Kobe, Japan) flow cytometer makes urine particle counting more precise and effective.
Several guidelines recommended that negative results of WBC or RBC from urine strips can be reported without checking by manual microscopy, and optimizing workflow also helps to decrease the microscopic examination of the samples with positive results of urine strips. 1   The reference procedure of the International Society of Laboratory Hematology (ISLH) recommended using native urine for the quantitative urine particle analysis. 8 However, if the amount of the clinically significant elements was too small in the urine, they may be missed.
Although we can use the same unit (cells/μL) to evaluate the cells in both centrifugal and non-centrifugal urine, the results of the same sample on these two conditions and clinical significance are obviously different. Centrifugation and discarding steps are the main error source of manual microscopy examination. Although the efficiency of centrifugation can be increased significantly by tripling the centrifugation, there is an increase in the particle destruction rate at the same time. Therefore, the particles in the supernatant do not mean that the number of particles in the sediment is decreased. Taken this into account, it is not surprised that comparison experiments showed that the UF series urine flow cytometer (Sysmex) may detect more RBCs and WBCs than microscopic examination did. 9 This finding brought confusion to clinicians in hospital which has multicenter laboratories applying both urinary sediment microscopic examination and flow cytometry analyzer.
Simultaneously utilizing the UF-1000i analyzer, FAST-READ chamber in our laboratory has already confused clinicians in our hospital.
The converting factor (0.297) we introduced here makes the original counts per μL (non-centrifugal urine) transferred to "counts per μL (equal to centrifugal urine)," and enables the urine particle analysis results of Sysmex UF-1000i correlated with that from chamber counting. The intercept of the regression lines found with the Passing-Bablok analysis is a negative value, which means that the UF-1000i method (after conversion) may have bias when the concentration of WBC or RBC is <1 cells/µL. The slopes of those regression lines, which are really close to 1, showed after conversion UF-1000i counting is comparable to the reference method.
According to Hannemann-Pohl et al, 6 the WBC count in native urine measured by UF series analyzer is 3.15 times of that derived from urine sediment analysis, which is very close to our converting factor (1/0.297 = 3.37).
Since the concentration of urine particles needs to be continuously monitored, it is better to keep the same reporting form for urine particle result. Thus, using the same reference intervals in a certain laboratory is more convenient.

ACK N OWLED G M ENT
We would like to thank Dr. Hongjun Zhu from Sysmex Corporation (Shanghai) for her assistance and guidance in this research.