| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Central Laboratory of Haifa and Western Galilee, Clalit Health Services
aaddress correspondence to this author at: Hematology Laboratories, Central Laboratory of Haifa and Western Galilee, Clalit Health Services, Nesher, Israel; fax 972-4-8209094, e-mail paulfr{at}clalit.org.il
Although examination of urine for blood hemoglobin by automated analysis of test strips is a common, precise, and efficient method with a maximum throughput of up to 300 test strips per hour (1), the reference interval for this procedure is uncertain. Most laboratories consider only a negative result as normal, based on findings in referred patients and on population studies showing that urinary tract tumors and other serious diseases are found in those with even trace amounts of blood on test strips (2)(3)(4). In two population studies, test strips were positive for blood in 3% of men of all ages (3) and in 13% of those over 60 years (4). A few patients with blood in their urine were found to have urothelial malignancies, but most had only trace results on initial testing, with re-testing yielding intermittently positive results (3)(4); intermittently positive results may be attributable in part to variations from visual inspection by technicians and by patients themselves which is less sensitive and less precise than semiautomated reflectance readings of urinalysis dipsticks (5).
Our laboratory receives complaints from physicians about the high frequency of hematuria in their patients. We are unaware of previous studies reporting the distribution of results in a healthy population tested by semiautomated reflectance readings. Recent studies using a urine flow cytometer (UF-100) suggest that the number of RBCs in the urine in healthy individuals should be in the range of 10–20 cells/µL with even higher estimated counts on dipsticks, which also measure lysed cells (6)(7)(8). In the present study we analyzed the urine of 1000 men and women of various ages seen consecutively for screening examinations and used automated analysis of the test strips to determine the reference interval.
Midstream urine samples from outpatient clinics arrived and were tested within 4 h of collection at a regional laboratory in northern Israel. Consecutive samples (n = 1000) from occupational medicine clinics that routinely perform urinalysis of asymptomatic workers were tested by use of a Supertron automated analyzer (Roche Diagnostics Ltd.) with Combur-10 S strips (Roche Diagnostics Ltd.). The patients were all actively working without chronic renal failure or symptomatic diseases and had no significant exposure to chemicals or dyes. We have shown previously that the Supertron test for erythrocytes is precise and that samples are stable over a 24-h storage period (1). We tested the accuracy of the assay by adding known numbers of erythrocytes to urine specimens. A logistic regression model was used to adjust for the presence of urinary glucose, nitrite, or protein and to determine the independent influence of age and sex on the risk of having 
50 erythrocytes/µL on urinalysis.
After excluding those with leukocyturia, glucosuria, proteinuria, or a positive nitrite result, we were left with 580 of the 744 men (78.0%) and 132 of 256 women (51.6%). After these exclusions the distribution of erythrocytes concentrations in urine for males changed little from that for the entire group (results not shown), whereas after the exclusions a lower proportion of females had 50 erythrocytes/µL [9 of 132 (6.8%) vs 19 of 124 (15.3%); P <0.05]. The optimum reference interval (95% of the individuals studied, after excluding those with other positive findings on test strips) for erythrocytes in urine for men <40 years of age included values up to 25 cells/µL, whereas for men 40 years and for females of all ages, the reference interval included values up to 50 cells/µL (Table 1
). On logistic regression, after we forced glucosuria (yes/no), nitrite (positive/negative), leukocyte esterase (yes/no), and proteinuria 300 mg/L (yes/no) into the model, the odds ratio for 50 erythrocytes/µL in females compared with males was 1.90 (1.02–3.55), the odds ratio for individuals with leukocyte esterase vs those without leukocyte esterase was 2.56 (1.35–4.83), and the odds ratio for individuals 40 years of age compared with younger individuals was 1.93 (1.08–2.45). We found no significant interaction between age and sex.
|
We added a known number of erythrocytes (counted on an Advia 120 Hematology System; Bayer) to urine to give a final concentration in the range of 6.25–400 cells/µL. The results by dipstick determination were nearly identical to the expected values [expected (test strip determinations): 0 cells added (12 negative tests by dipstick), 6.25 cells/µL added (0, 10, 10, 0, and 0 cells/µL by test strip); 12.5 cells/µL added (10, 10, 10, 25, and 10 cells/µL by test strip); 25 cells/µL added (25, 25, 25, 25, and 25 cells/µL by test strip); 50 cells/µL added (25, 50, 50, 50, and 50 cells/µL by disptick)]. Test strip results were also highly correlated with final concentrations of 100, 200, and 400 cells/µL (results not shown).
The major finding of this study is that the reference interval for erythrocytes in urine should include positive results. We have also shown that after calibration, test strip results are accurate over various known concentrations of erythrocytes in urine. The reference interval was age dependent, increasing with age, as has been reported previously with a high-power-field methodology (9). Our results can probably be extrapolated to other analyzers that use similar technologies after calibration with known concentrations of erythrocytes, a necessary step because adequate commercial controls are not available.
Our semiquantitative results are consistent with those obtained with a urine flow cytometer in a small number of apparently healthy children and adults. In young men, we found that the upper limit of the reference interval should be 25 cells/µL erythrocytes, whereas Lun et al. (6), using the UF-100 urine flow cytometer, reported that the 95th percentile for erythrocytes in 141 children was 25.9 cells/µL. Regeniter et al. (7) tested 91 healthy individuals (both males and females; age range, 11–63 years) but could not consider subgroups because of the small sample size. They reported 97.5% percentiles of 13.9/µL for erythrocytes and 15.7/µL for eukocytes; they also reported that the overall mean UF-100 erythrocyte and leukocyte counts were somewhat lower than the results obtained with the semiquantitative stick method, suggesting that lysed cells contributed to the higher observed values. Recently, Marimoto et al. (8) reported median erythrocyte counts on a flow cytometer of 
10 cells/µL in 64 healthy female students 18–20 years of age. The 90th percentile upper limit ranged from 50 to 600, depending on whether the students were menstruating. The 90th percentile range in our females included 50 cells/µL, which is consistent with the results obtained for nonmenstruating students in that study. The automated quantitative urinalysis analyzer (UF-100) cannot be considered a gold standard because the imprecision of the erythrocyte count for this analyzer is high [CV, 18–31% for 5–60 cells/µL (10)]. Furthermore, such testing does not include lysed erythrocytes.
A major strength of our study is the large number of screening tests done in apparently healthy individuals. Although we did not have complete medical histories available to us, it is unlikely that concomitant diseases biased our results. All participants were actively employed and received medical clearance to work. Chronic diseases are more prevalent in those 50–59 years of age compared with those 40–49 years of age, but the frequency distribution of erythrocytes in urine was nearly identical in the two groups. After exclusion of those with glucosuria, proteinuria, leukocyturia, or nitrites in the urine, the reference intervals were unchanged. Finally, because the prevalence of microhematuria is not affected by physical exercise in the 24 h preceding urinalysis in asymptomatic young men (11), leisure physical activity probably did not significantly affect our results.
Standard textbooks do not consider semiquantitative or quantitative urinary erythrocyte analysis but instead continue to recommend that a complete workup be done for any patient with a persistent finding of >3 erythrocytes per high-power field on urinalysis, regardless of the pretest probability of significant urothelial disease (12)(13). Recent recommendations support a repeat urinalysis after 48 h to confirm positive test strip findings and a work-up that includes confirmation of the dipstick results by microscopic urinalysis (14)(15) and identification of dysmorphic erythrocytes (14)(15). This approach decreases the sensitivity for detecting serious urothelial diseases because of the intermittent nature of hematuria in patients with serious urothelial disease (2)(3)(4)(16), the known problems with the precision and accuracy of microscopic urinalysis (2)(17)(18), with intralaboratory CV of 25–50% (8)(19), and with the possibility of concomitant dysmorphic erythrocytes in the urine in the presence of urothelial cancer (20). It has been pointed out that current data are inadequate to support clear-cut recommendations regarding the management of microscopic hematuria (15).
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
