Dipstick Testing of Urine Can It Replace Urine Microscopy?

Transcription

1 Dipstick Testing of Urine Can It Replace Urine Microscopy? MARY C. MORRISON, M.D. AND GIFFORD LUM, M.D. One thousand consecutive urine specimens were studied to assess the sensitivity of a commercially available dipstick (Chemstrip 8, Boehringer Mannheim Corp., Indianapolis, IN) to predict the presence or absence of microscopic abnormalities. The Chemstrip 8 had a sensitivity of 78%, specificity of 54%, and a false negative rate of 38%. An additional consecutive urine specimens were then studied using the Chemstrip 9, a reagent dipstick that includes the leukocyte esterase (LE) test. The Chemstrip 9 had a sensitivity of 82%, specificity of 42%, and a false negative rate of 36%. Chi-squared analysis revealed that the two dipsticks were not significantly different (x 2 = 0.17, P > 0.5). Clinical review of patients with false negative results showed that approximately one-third to one-half of these patients had either spinal cord injury or genitourinary problems. Maximal potential savings in workload of approximately 10% were found if microscopic examinations were to be performed only on urine specimens with abnormal dipsticks. Our data suggest that in our patient population, we should not eliminate microscopic urine examination based on abnormal dipstick findings. (Key words: Urinalysis; Dipstick screening; Urine microscopy) Am J Clin Pathol 1986; 85: TRADITIONALLY, urinalysis and microscopic examination of the urinary sediment have provided valuable clinical information. Physicians have relied upon these laboratory procedures for a number of applications, including screening for diagnosis of asymptomatic disease, monitoring, and following the course of a disease. Over the years, physicians have come to rely increasingly on the laboratory to provide these tests. Thus, urinalysis has evolved from an essentially bedside procedure to a wellaccepted, standardized laboratory procedure, consisting of a chemical screening dipstick and microscopic urinary examination. Unfortunately, because of the routine nature of this test, technologists often spend comparable time examining both normal and abnormal urine specimens in an indiscriminate fashion. Because routine urine microscopy is labor intensive and subject to less careful scrutiny if done by rote, we evaluated the feasibility of prescreening urine specimens with a chemical dipstick prior to microscopic analysis. We hoped Laboratory Service, Brockton/West Roxbury Veterans Administration Medical Center, and Harvard Medical School, Boston, Massachusetts to achieve savings in technician time and a superior microscopic examination. We therefore undertook a prospective study of consecutive urine specimens using, without modification, our current urinalysis protocol to assess the effectiveness of prescreening urine specimens to select those requiring microscopic examination. The hypothesis that we wished to test was the following: using this urine screening protocol, we could effectively select those specimens requiring versus those not requiring microscopic examination on the basis of dipstick findings. We then studied an additional consecutive urine specimens using a modification of our current urinalysis protocol to include leukocyte esterase (LE) testing to assess the impact of this additional test to identify microscopically abnormal urine samples. The hypothesis that we wished to test was the following: addition of LE testing to the dipstick would improve the effectiveness of selection of those specimens requiring versus those not requiring microscopic examination. Dipstick Analysis Materials and Methods and 9 reagent strips were obtained from Boehringer Mannheim Corporation, Indianapolis, Indiana, and used according to the manufacturer's specific instructions. Chemstrip 8 includes reagent pads for assessment of ph, protein, glucose, ketones, nitrite, urobilinogen, bilirubin, and hemoglobin. Chemstrip 9 includes all reagent pads present in the Chemstrip 8 as well as a reagent pad for semiquantitation of LE activity. Received June 18, 1985; received revised manuscript and accepted for publication August 22, Address reprint requests to Dr. Lum: Brockton/West Roxbury Veterans Administration Medical Center, 1400 VFW Parkway, Boston, Massachusetts Visual Analysis Visual examination consisted of an assessment of color and turbidity by an experienced laboratory technologist. 590

2 vol. 85-No. 5 URINE DIPSTICK TESTING 591 Table 1. Comparison of Urine Results Obtained Using and 9 with Examination Part A Chemstrip 8 PartB Chemstrip 9 Dipstick Sensitivity = 78% Specificity = 54% Dipstick Sensitivity = 82% Specificity = 42% Analysis of Urinary Sediment Urine specimens were prepared for microscopic examination by centrifuging 20 ml of urine at rpm for 5 minutes. The supernate was decanted, and approximately 0.05 ml of sediment was placed on a KOVA urinalysis slide tray (KOVA, Fountain Valley, CA). The unstained urinary sediment was then examined at low (X10) and at high (X40) power for the presence of formed elements, including white blood cells (WBCs), red blood cells (RBCs) bacteria, crystals, fungi, casts, renal tubular cells, epithelial cells, and viral inclusions. Whenever possible, the number of elements per high-power field (hpf) was approximated. Urine Samples One thousand consecutive urine samples from hospitalized and clinic patients were prescreened initially with the Chemstrip 8 and visual inspection. This patient population was predominately adult males (approximately 96%, ranging in age from years old). Urine samples greater than 24 hours old were excluded from the study. Specimens were classified as physicochemically normal based on the following criteria: color clear yellow; ph 5 to 7; glucose, ketones, bilirubin, urobilinogen trace or negative; protein, hemoglobin and nitrite negative. Urine specimens not meeting these criteria were classified as physicochemically abnormal. Following physicochemical analysis, all urine specimens were then examined microscopically. Urine specimens were classified as microscopically normal based on the following criteria: <6 WBCs, RBCs, or renal tubular cells/hpf; <2 bacteria; and no casts, pathologic crystals (such as cystine, tyrosine, and leucine), fungi, parasites, or viral inclusions. Urine specimens that did not meet these criteria were classified as microscopically abnormal. We then screened another consecutive urine samples from patients with the Chemstrip 9 and visual inspection. This patient population was predominately adult males (approximately 97%, ranging in age from years old). Criteria adopted for classification of urine samples into physicochemically normal/abnormal categories were as outlined for the Chemstrip 8 with the addition of LE test. If the LE reagent strips showed any reaction (trace, 1, 2,...) at 60 to 120 seconds, the specimens were classified as physicochemically abnormal. Following Chemstrip 9 screening, all urine specimens were examined microscopically, and criteria adopted for classification of samples as normal/abnormal were the same as for the first part of this study. The microscopic examination of urinary sediment was adopted as a standard of reference in order to assess the effectiveness of our two screening protocols. After urine specimens were assigned to physicochemical and microscopic normal/abnormal categories, respectively, the results were summarized in the following 2X2 matrix: Dipstick ( or positive) (- or negative) ( or positive) ( or negative) Results Table 1 (Part A) displays the data obtained for consecutive urine specimens that were screened using the Chemstrip 8. There were 658 urines with abnormal dipsticks and of these, 28% (false positive, 182/658) showed no microscopic abnormalities. On the other hand, there were 342 urine specimens with normal dipsticks, and of these, 38% (false negative rate, 131/342) did show microscopic abnormalities; these included, listed in order of decreasing frequency, bacteria (63%), WBCs (42%), casts (24%), RBCs (5%), and yeast (2%). Specimens frequently

3 592 MORRISON AND LUM A.J.C.P. May 1986 Table 2. Distribution of Clinical Diagnoses in Patients with False Negative Physicochemical Screening Tests Group 1 (spinal cord injury) Group 2 (GU problems) Group 3 (other) Part A (n = 105) Percent Part B (n = = 69) Percent contained more than a single abnormality and were counted more than once, resulting in total percentage exceeding 100%. Table 1 (Part B) displays the data obtained for the consecutive urine specimens that were screened using the Chemstrip 9. There were 721 urines with abnormal dipsticks and of these, 34% (false positive rate, 245/721) showed no microscopic abnormalities. On the other hand, there were 279 urine specimens with normal dipsticks and of these, 36% (false negative rate, 101/279) did show microscopic abnormalities; these included, listed in order of decreasing frequency, bacteria (84%), WBCs (26%), and casts (10%). Again, specimens often contained more than a single abnormality and were tallied more than once, resulting in a total percentage exceeding 100%. In order to ascertain whether or not the Chemstrip 9 performed significantly better than the Chemstrip 8 in the detection of urinary microscopic abnormalities, we analyzed the following data for all negative dipstick results: False negatives True negatives All Negatives Chemstrip 8 Chemstrip The false negative rates for the Chemstrips 8 and 9 were not significantly different (chi-square analysis = x 2 = 2 t( ~ e ) ~ J = o.l7, P > 0.5). There was no statise tically significant difference in the proportion of false negative results for either dipstick The false negative urine specimens represent those specimens in which microscopic abnormalities would be undetected using dipstick screening. We therefore reviewed the medical records of those patients with false negative results and abstracted the following clinical information: age, sex, hospital status, and clinical diagnoses. Because of the wide spectrum of specific clinical diagnoses, we devised a simplified classification scheme by categorizing patients into three broad groups. This grouping procedure allowed us to make rational judgements concerning the diagnostic groups in which dipstick screening may not be effective. Group 1 included patients with spinal cord injuries who were either quadriplegic or paraplegic and no voluntary urinary control. Group 2 included patients with genitourinary (GU) problems, whether current or past e.g., transurethral prostatic resection, prostatic hypertrophy, nephrolithiasis, cystic calculi, and testicular leiomyomata. Group 3 included all patients who did not fall into the genitourinary problem category or the spinal cord injury group. Some of the more frequent diagnoses in this group included rheumatoid arthritis, coronary artery disease, hypertension, diabetes mellitus, strokes, and chronic alcoholism. Recognizing that there may be some overlap between the groups, e.g., a spinal cord injury patient with general medical problems or a patient with genitourinary problems and cardiovascular disease, we arbitrarily decided to assign all patients to one group only and adopted the following algorithm: All patients Assigned Category Spinal cord injury patient present, ± any other diag- Group 1 nosis, including genitourinary No spinal cord injury, genitourinary problems Group 2 present ± any other diagnoses t = Px -P-, (1-P,) 2 ( (1 -P 2 ) 2 = 0.49 ni where p = % false negative. n 2 No spinal cord injury, no genitourinary fr Group 3 problems, any other diagnoses

4 Vol. 85-No. 5 URINE DIPSTICK TESTING Table 3. Workload Comparison Between Dipstick Screening and No Dipstick Screening* 593 Dipstick Screening No Dipstick Screening Workload 342 1, , ,316 1, ,326 5,442 6,000 6,000 * Urinalysis without microscopic examination = 4 CAP units. Urinalysis with microscopic examination = 6 CAP units. Potential savings (Chemstrip 8) = 11.4% (684/6,000). Potential savings (Chemstrip 9) = 9.3% (558/6,000). For the first part of our study involving the Chemstrip 8 screening protocol, there were 131 urine samples that were physicochemically normal but showed microscopic abnormalities. These false negative samples represented 111 individual patients because multiple urine samples were sometimes submitted for the same patient. We were able to locate 105 records; the six patients whose medical records were hot found were excluded from this review. Table 2 (Part A) shows that patients with spinal cord injury and genitourinary problems accounted for approximately 50% of the false negative results using Chemstrip 8. For the second part of our study involving the Chemstrip 9 screening protocol, there were 101 urine samples that were physicochemically normal but showed microscopic abnormalities. These false negative samples represented 83 individual patients. We were able to locate 69 patient records; the 14 patients whose medical records were not found were excluded from this review. Table 2 (Part B) shows that patients with spinal cord injury and genitourinary problems accounted for approximately 35% of false negative results using Chemstrip 9. College of American Pathologists (CAP) workload units can be applied to measure the impact of using the conventional procedure (complete urinalysis) versus adopting a screening protocol. Table 3 shows a workload comparison between Chemstrip 8 and 9 screening versus no screening. Potential savings in workload were 11.4% and 9.3% for the Chemstrip 8 and 9, respectively. The "cost" for finding a single microscopic abnormality that would otherwise be undetected (normal dipstick, no microscopic), would be 5.2 and 5.5 CAP units for Chemstrip 8 and 9, respectively, derived by dividing the potential savings in workload by the number of false negatives in each category. Discussion Much current interest has focused on appropriate utilization of scarce laboratory resources in the present-day cost-conscious environment, and urinalysis has been a paradigm of many recent investigations. Some authors feel that workload can be reduced significantly and clinical standards of excellence maintained by prescreening urine samples using a dipstick and only performing microscopic urinalysis on those urine samples showing abnormal biochemical dipsticks. 1 " 5,8 " 12,1415 In contrast, other authors feel that prescreening urine specimens using a dipstick and only performing urinalysis on physicochemically abnormal urines is of little value because the dollar savings are outweighed by the loss of potentially valuable clinical information gained by this low-cost screening test. 6 ' 713 This study was undertaken to determine whether the laboratory could reliably predict, on the basis of a physicochemical screening test, the urines that would show microscopic abnormalities. The aim of the study was twofold: first, using procedures already in routine use in nearly all clinical laboratories, can the laboratory justify a simple screening urinalysis without microscopic examination? And second, does the addition of LE testing, widely regarded to optimize the detection of WBCs in urine, 110 improve the ability of the laboratory to screen for urinary microscopic abnormalities? In our hands, neither the Chemstrip 8 nor the Chemstrip 9 is a reliable predictor of abnormal urine microscopy. This is reflected in the high false negative rates seen in our patient population. The differences found between our study and the studies of those who advocate the use of dipstick screening may reflect the high proportion of patients with spinal cord injury and with known urinary tract disease represented at our institution. Elimination of the microscopic examination, in our experience, resulted in only marginal workload savings (11.4 and 9.3%). We feel that this rather modest savings is overshadowed by the potential loss in valuable clinical information. The inclusion of the microscopic examination resulted in detection of abnormalities at a modest cost in technician time. The yield of finding abnormalities is high compared with the small cost in time. Based on the unacceptably high false negative rate

5 594 MORRISON AND LUM A.J.C.P. May 1986 found in our study, we plan to continue performing urine microscopic examination despite dipstick findings. We caution those laboratories contemplating elimination of routine urine microscopy based on the dipstick screen to consider whether such screening would be effective in their patient population. References 1. Bartlett RC, Kaczmarczyk LA: Usefulness of microscopic examination in urinalysis. Am J Clin Pathol 1984; 82: Christenson RH, Tucker JA, Allen E: Results of dipstick tests, visual inspection, microscopic examination of urine sediment, and microbiological cultures of urine compared for simplifying urinalysis. ClinChem 1985;31: Freedman SI: Routine microscopic examination of the urine sediment revisited. Arch Pathol Lab Med 1984; 108: Gelbart SM, Chen WT, Reid R: Clinical trial of leukocyte test strips in routine use. Clin Chem 1983; 29: Kiechle FL, Karcher RE, Epstein E: Routine microscopic examination of the urine sediment revisited. Arch Pathol Lab Med 1984; 108: Nanji AA, Adam W, Campbell DJ: Routine microscopic examination of the urine sediment Should we continue? Arch Pathol Lab Med 1984; 108: Resnick M: Simplifying urinalysis. Clin Chem 1985; 31: Sarewitz SJ: Routine microscopic examination of the urine sediment revisited. Arch Pathol Lab Med 1984; 108: Schumann GB, Greenberg NF: Usefulness of macroscopic urinalysis as a screening procedure A preliminary report. Am J Clin Pathol 1979;71: Shaw ST, Poon SY, Wong ET: Routine urinalysis Is the dipstick enough? JAMA 1985; 253: Smalley DL, Bryan JA: Comparative evaluation of biochemical and microscopic urinalysis. Am J Med Technol 1983; 49: Speicher CE: Routine microscopic examination of the urine sediment revisited. Arch Pathol Lab Med 1984; 108: Szwed JJ, Schaust C: The importance of microscopic examination of the urinary sediment. Am J Med Technol 1982; 48: Valenstein PN, Koepke JA: Unnecessary microscopy in routine urinalysis. Am J Clin Pathol 1984; 82: Valenstein PN, Koepke JA: Routine microscopic examination of the urine sediment revisited. Arch Pathol Lab Med 1984; 108: