Feature Story

Urinalysis finally catches up

September 2002
William Check, PhD

Sometimes asking a laboratorian what is changing in the field of urinalysis produces a brief silence, followed by a remark such as, "Urinalysis has kind of stayed urinalysis," or "Urinalysis is not the most exciting and dynamic area in laboratory medicine." But a minute of thought usually brings to mind important efforts being made to improve the value of urinalysis to ordering physicians and patients. These efforts range from assessing when urinalysis should be ordered and how results can be made more reliable, to the introduction of automation into microscopic analysis of urine, one of the few areas of laboratory medicine that is still done manually in many hospitals.

"We can look at urinalysis from two different viewpoints," says William Rock, MD, professor of pathology and director of the clinical laboratory at the University of Mississippi Medical Center in Jackson. In the first and most common circumstance, routine screening, "Results of urinalysis are often not optimally taken advantage of by clinicians," Dr. Rock says. His biggest concern is that an important result will be missed. "When urinalysis is part of a routine situation, physicians expect it to be negative," Dr. Rock says. Because of this expectation, "alertness is lowered and the results don’t get as much attention." A meaningful step, then, in improving the use of urinalysis is to encourage physicians to review the results of urinalysis every time they order one.

In the second circumstance, urinalysis is requested because the patient has signs and symptoms suggesting that something is wrong with the urinary tract. In that case, urinalysis becomes more important.

Nephrologist Giovanni Fogazzi, MD, who directs the laboratory of the renal unit at the Ospedale Maggiore of Milan, carries this line of thought further. "Urinalysis is not generally ordered appropriately," he says. "It is a test that is abused. If it is done just for screening, nobody cares. We know now that urinalysis is not for everybody, but should be requested only on the basis of a clinical problem, especially for the kidneys." The position that urinalysis should be ordered selectively is incorporated into the latest ECLM-European Urinalysis Guidelines, which Dr. Fogazzi helped formulate (Scand J Clin Lab Invest. 2000;60[suppl 231]; www.pshp.fi/labnet/EUgroup.htm).

Improving the quality of urinalysis results is another concern, says Timo Kouri, MD, PhD, associate chief physician at Oulu University Hospital, Finland, and chairman of the European project that produced the guidelines. Dr. Kouri notes that the computer maxim "Garbage in, garbage out" applies as well to urinalysis. "People have been putting in junk and then throwing up their hands because they don’t know the relevance of the results that come out," he says. Improving the quality of urinalysis results means improving every step of the process, starting from the preanalytical phase, including specimen collection and transport. "Every detail should be outlined within a standardized operating procedure," Dr. Kouri says.

The most dramatic advance in urinalysis has been the evolution of automated instruments to replace visual microscopic analysis of formed elements in urine. "This has been the biggest change in urinalysis in the last 20 years," says Edward Fody, MD, chief of pathology at Erlanger Medical Center in Chattanooga, Tenn. "The coming of automated particle analysis systems has revolutionized urinalysis."

Adds Dr. Kouri, "Automated analysis and counting of cells or particles in urine, which may also include casts or bacteria, saves time and improves precision of counting."

Gifford Lum, MD, assistant chief of Pathology and Laboratory Medicine Service at the VA Boston Healthcare System, notes that the microscopic examination of urine is one of the last areas of the laboratory to be automated. "The problem has always been that manual microscopic examination of urine was tedious, so it was not done uniformly," he says. He compares the situation to cell counting in hematology. "Before the coming of automated hematology instruments, there were protocols not to do differentials on all CBCs," he says. Now that has changed. "We will see the same thing happening with urinalysis," Dr. Lum predicts. "When it gets automated with flow cytometry, it will be done automatically regardless of the results of dipstick tests."

"Urines have languished," says Jeri Walters, MT, technical supervisor at ACL Laboratories, Milwaukee. "Now urinalysis is finally catching up with the rest of the laboratory in technology and automation. The latest technology operates much like an automated hematology instrument," says Walters, referring to the Sysmex UF-100, a flow cytometer that performs automated analysis of formed elements in urine. "It works much like our five-part differential analyzer, doing screens and throwing flags for abnormalities." Walters finds that automated screening for cells and particles affects attitudes as well as accuracy. "It changes how you think about results," she says. "They are no longer ’crummy close-enough urine results.’"

Speaking of the other major automated instruments for analyzing formed elements in urine, from IRIS Diagnostics, Dr. Rock says, "This instrument has enabled us to automate and operate full time."

Bacterial culture, the third part of a complete examination of urine, "seems to be the most common microbiology laboratory test in the world," says Dr. Kouri. It used to be routine for every patient with symptoms possibly attributable to urinary tract infection, he says. But the new European guidelines classify low- and high-risk patients and recommend routine cultures only for high-risk cases.

In the new European guidelines, emphasis is on "medical need" rather than general screening. "In some parts of the world, people think of urinalysis to be done for everybody," Dr. Kouri says. "Now we think first of medical need, then cost versus benefit." For patients with diabetes or hypertension, for example, a sensitive analysis detecting proteinuria or albuminuria is becoming more frequent. "For other patients, it may not be justified to study urine at all," Dr. Kouri says.

The state of public health care is another consideration. In Finland every child is examined by a public health doctor annually, Dr. Kouri says. In Japan, a child may not have an annual examination unless the parents organize it. "In that case, urine screening may be a replacement for a physician," he postulates.

The situation at Chicago’s Stroger Hospital of Cook County may be similar. According to Mariano Yogore, MD, director of the hospital’s Adult Emergency Services Testing Center, the hospital has one of the largest emergency services in the nation, serving 120,000 patients each year. About 45,000 of these patients have urinalyses. "I think there is some overuse," Dr. Yogore says. "Because of our volume, patients are first seen by triage nurses or physician assistants, who may order urinalysis as soon as a patient is interviewed and examined." Also, many patients are economically disadvantaged and non-English speaking. In this setting, urinalysis may be used as a screening examination or in lieu of a medical history.

In Italy, urinalysis is one of the most commonly requested tests, says Dr. Fogazzi, who is the author of a monograph on the urinary sediment (Fogazzi GB, Ponticelli C, Ritz E. The Urinary Sediment. An Integrated View. 2nd ed. Oxford: Oxford University Press; 1999). It is ordered at every office visit and every hospitalization. "This has practical and important consequences," he says. "In my last workshop, three or four of the 10 physicians attending had to do microscopic sediment analysis every day on at least 300 urine sediments. This implies inevitably poor quality and unreliable results."

Under the best of circumstances, physicians are given a lot of information and urinalysis is just one more small piece, says Dr. Rock. He cites an article in JAMA about 10 years ago that evaluated clinicians’ performance in reviewing the medical chart. "Urinalysis was the least looked at item on the medical report," he recalls. To Dr. Rock, this does not prove that urinalysis is overordered for screening. "If there was ever a test that you could justify for screening, urinalysis is it," he asserts. In his view, it quickly provides an assessment of a variety of functions. "If it is normal or negative, we have eliminated certain diseases from primary consideration," he says. "But," he reiterates, "you have to look at it and pay attention to the results."

In the quest for better urinalysis results, written or clear-cut oral instructions to the patient on how to collect a urine sample is the first step. "Here in our unit we have written instructions," says Dr. Fogazzi. "They could be improved, but it is a start." Frank illustrations on how to collect a midstream urine are available in many languages on the same Web site as the European guidelines. "It is now possible to instruct the patient and afterward ask whether there were any problems in specimen collection," Dr. Kouri says. In Finland, specimens are divided into two categories based on the answer—qualified or non-qualified or unknown.

"There are standards for a midstream urine," acknowledges Dr. Rock, "but it is very hard to get." To avoid contamination, he recommends processing urine specimens in two hours and refrigerating if necessary. "We use the criterion of three or more epithelial cells/high power field," he says. If this threshold is exceeded, the results are labeled "Interpret with caution." "We routinely report the number of epithelial cells we see," he adds.

Walters agrees: "Getting clean-catch midstream samples is always going to be a problem. Someone might give the patient a piece of paper with instructions and assume they understand our technical terms. Or someone explains it in 20 seconds." Health care providers should be aware that they need to instruct patients better.

Dr. Yogore calls proper specimen collection "one of most important things that has to be followed in an ER setting, where we are overwhelmed with patients." Unfortunately, he says, "Many patients in the population we serve may not understand the instructions." Diagrams are provided at the nurses’ station but may not be sufficient. "Contamination due to improper collection, especially with females, is an ongoing problem," Dr. Yogore says. "It is reflected in the need for repeat specimens."

Another step in improving urinalysis results is to centralize specimen collection and improve the transportation chain. In large countries—the U.S., Canada, Russia, Australia, parts of France, the United Kingdom—this requires effective preservation with either boric acid or mercury-chloride-based preservatives, Dr. Kouri says. He calls mercury chloride, often used in the U.S., "not very customer friendly" and "toxic." Dr. Kouri published this year in Clinical Chemistry (2002; 48: 900-905) an evaluation of a commercial boric-acid-based solution showing that it preserves bacterial counts for two to three days as well as leukocytes and particles for visual microscopy and automated analysis. "In the future," he predicts, "development will again go toward preservatives for bacterial culture and test strips and particle analysis because of centralization of health care and lack of laboratory technologists all over the world."

Walters favors Becton Dickinson preservative tubes for samples not analyzed within one hour of collection, such as outreach work. She uses either a tube with a camouflage top and conical bottom, designed for routine urinalysis, or a gray-top tube, designed for microbiology. Gray-top tubes preserve urine elements well, Walters finds. "The biggest shift we saw [with gray-top tubes] was a slight elevation in specific gravity," she says. "Shifts in other chemical components were minor compared to leaving a specimen sit for 12 hours."

As with many other changes in laboratory medicine, enthusiasm for automated analysis of formed elements in urine is driven to a large extent by economic constraints and the shortage of medical technologists. In this climate, microscopic analysis of urine tends to be restricted. "Like a lot of laboratories at a time of decreasing staff and increasing pressures to perform, we were following a protocol," says Leo Serrano, FACHE, CLSup (NCA), executive director of laboratory services at West Tennessee Healthcare, Jackson. Microscopic examination was not done if dipstick results were all within normal limits and the urine was clear, unless the ordering physician requested it. "Even so, we were doing more than 200 urine microscopies every day," Serrano says. "So we started looking for an automated method."

Two approaches dominate automated particle analysis. Sysmex instruments are flow cytometers, using forward scatter to classify particles by size and two fluorochromes—one that stains membranes and the other nucleic acids—to differentiate cells or casts based on their contents.

IRIS Diagnostics’ instruments route samples through a flow cell where a digital camera captures images of each particle. Images are displayed on a computer screen and an operator identifies and classifies each element by touching the screen. A newer IRIS instrument, the 900 series, filters images through neural network-based pattern recognition software, which automatically classifies elements based on morphological characteristics.

Dr. Kouri has published a validation study comparing the accuracy of the Sysmex UF-100 to visual microscopy. He found that the instrument improved precision and accuracy and saved time and labor. It was equal to microscopy for red blood cells, white blood cells, and bacteria, which is what clinicians want to know about in most cases (Am J Clin Pathol. 1999;112:25-35). A study conducted by Professor Joris Delanghe of the University of Ghent (Clin Chem. 1999;45:118-122) corroborated Dr. Kouri’s results.

Validation data on the IRIS 400 series were presented in abstract form at 1991 (Abs #91) and 1994 (Abs #58) ASCP/CAP meetings. This work showed that the IRIS was at least as sensitive as standard manual methods for most abnormalities and possibly more sensitive for some.

From the point of view of nephrologists, however, automated instruments are not helpful, Dr. Fogazzi notes. "These instruments don’t reliably distinguish subtypes of casts and do not discriminate tubular from transitional—superficial and deep—cells," he says. "In addition, they don’t recognize lipid particles at all, which are a frequent finding in patients with nephrotic syndrome and may be diagnostic for lipid storage disorders such as Fabry’s disease." So a nephrologist would have to look at all specimens that have any particles. "But for general laboratories, where several hundreds of sediments are analyzed every day, and most sediments are normal and only a fraction contain abnormal findings, to have an instrument like that is useful," he says.

Dr. Fody bought a Yellow IRIS in 1984 when he was at the Little Rock, Ark., VA hospital. It produced reliable results and turned a workstation where none of the medical technologists wanted to work into "a very desirable workstation," he says. Earlier this year he purchased a Model 500 IRIS for his lab at Erlanger Medical Center. "There is definitely a technologist interaction," he says, "and there always should be." A technologist has to verify the instrument’s judgments, but they find looking at the computer screen easier and faster than looking through a microscope.

Dr. Rock has two IRIS instruments at the University of Mississippi and is upgrading to the 500 series. "My technologists really like the IRIS and how it works for them," he says. The operator uses a pointer, finger, pencil eraser, or the like to touch each element on the screen for each category. "If the urine is negative, the process goes very quickly," he says. "If there are lots of formed elements, such as forms highly suspicious of casts or crystals, we would do a manual microscopic analysis." Manual microscopic analysis is necessary in from 10 to 20 percent of cases.

Whether the instrument is labor saving is not the key factor, in Dr. Rock’s view. "For us, the IRIS helped to define urinalysis as a workstation running 24/7 with one operator," he says. "It lends itself to late-night shifts where you have various levels of expertise." Dr. Rock’s laboratory processes 300 to 400 urine specimens daily. Every sample goes through both chemical and microscopic analysis. "We have a large sick population and physicians want to hear all negative microscopic results, particularly with oncology patients," he says.

Dr. Rock looked at an IRIS 900 series instrument, which is truly walkaway in terms of loading, though it still requires a technologist to look at images and verify classifications and do microscopy when indicated. "We’ll reassess the 900 as our volume increases," he says.

At Stroger Hospital of Cook County, a large, new hospital has just been built to replace the old one, and nearly all laboratories will be integrated into a new central hospital laboratory. This facility will have two IRIS 500 series instruments that will handle about 65,000 urinalyses annually. A high percentage will get a microscopic examination. Even though the IRIS 500 instrument does not batch process, Dr. Yogore is confident it can manage the high volume because of its reliability and speed.

"I myself have no experience with the IRIS," Dr. Yogore says. He chose this vendor because the outpatient laboratory at Stroger Hospital had been using an IRIS for several years. Most recently the outpatient laboratory got a 900 series, but it has had problems with the FESH unit, the front-end sample handler. "So," Dr. Yogore says, "even though the 900 series can do batch testing, we selected the -manual-loading 500 instrument. It is based on the 400 series, which has a long history as a workhorse."

IRIS instruments save images, which is highly useful. The instruments make it possible for the Stro-ger Hospital laboratory to compare interpretation among different operators and standardize reporting, Dr. Yogore says, which provides an excellent quality control mechanism and a teaching tool.

In the laboratory at the VA Boston Healthcare System, Dr. Lum has a Yellow IRIS that he inherited from another hospital when his laboratory became the tertiary care center for the health system. "My concern about the IRIS is that there is too much interaction between operator and instrument," he observes. "A technologist has to identify components such as WBCs, RBCs, casts, et cetera as they come up on photo images. So I don’t really think it has freed up any technical time or contributed to automation of urinalysis."

Flow cytometer-type instruments are the wave of the future, in Dr. Lum’s view. Current flow cytometer-based instruments, such as the Sysmex UF-100, are truly walkaway in terms of loading, he notes. "I think the UF-100 is probably the way to go," Dr. Lum says. He adds that flow cytometers can also do cell counts on several body fluids, such as cerebrospinal, ascitic, and peritoneal fluids.

When Serrano was deciding on an automated instrument for urine microscopies at West Tennessee Healthcare, he brought a Sysmex UF-100 into his laboratory and paralleled it with standard microscopic examinations in an internal study. He took 3,000 dipstick-negative urines and ran them through the UF-100. Anytime the UF-100 detected what was considered a clinical abnormality—more than 5 WBCs/ µL, more than 3 RBCs/ µL, any casts, etc.—the specimen was centrifuged and examined through the microscope. "Roughly 18 percent of the 3,000 dipstick-negative urines had some sort of formed element abnormality or the presence of some formed elements that we would not have reported previously," Serrano says. "That got our attention." The laboratory purchased a UF-100 about one year ago and now performs flow-based microscopic analysis on every urinalysis, regardless of dipstick results. "We have been very pleased with the findings," Serrano says.

With the automated instrument, they were able to reduce staffing in the area by one full-time equivalent. "Urinalysis microscopy used to be a full-time assignment on day shift and evening shift and part-time on night shift," Serrano says. Now it is an add-on assignment on each of those shifts for the same person who does manual analysis on differentials.

Originally Serrano set up the instrument so that flagged specimens would be reviewed manually. "Now that we have confidence," he says, "we set it up to autorelease most findings." Only certain flags require validation. "We still have a person look at fine and coarse granular casts and cellular formed casts," he says. "We trust the instrument to detect RBCs and WBCs. We visually validate yeast and yeast-like organisms." With this protocol, the number of microscopic examinations has been reduced by about 80 percent, Serrano estimates.

At ACL Laboratories, Walters chose between the two automated instruments for microscopic analysis in much the same way that several other laboratorians did—based on the experience of a colleague. ACL Laboratories is a large merged core laboratory and one of the merged sites had a positive experience doing beta site testing with the UF-100. "We were pretty confident there was nothing else on the market at this time with that level of technology and sophistication for urinalysis," she says. Previously she had had an IRIS 900 series instrument, but, she says, "that was still very labor intensive. It was better than manual, but still subject to technologist-to-technologist variation and fatigue."

After about eight months, she finds the UF-100 advantageous. "A big step up for urinalysis is the increase in precision you get out of this technology," she says. "It is more reproducible than manual methods, where there are too many variables." She finds that the UF-100 saves about one FTE. "It depends on patient acuity," she notes. "If you have a lot of clear, colorless urines, they will go through anything fast."

Walters has set up protocol to verify anything that has a possible pathological finding, such as increased renal cells, pathological casts, and the presence of yeast or sperm. "We will accept RBCs and WBCs from the instrument," she says.

One possible screening criterion her lab has studied is the crystal flag. "We weren’t sure if it was really significant," Walters says. So she compared detection of abnormal samples with all flags set except the crystal flag versus screening with the crystal flag activated. Among 610 samples, overall sensitivity for abnormal samples without the crystal flag was 96.9 percent and specificity was 89.8 percent, with a false-negative rate of three percent. (The only samples missed had yeasts without WBCs, suggesting contamination.) "So we really feel the crystal flag was not significant," Walters concludes. As a screening test, the negative clinical accuracy (negative predictive value) was 99.4 percent, which, Walters says, "is right where we are aiming."

In contrast to analysis of formed elements, which is being applied more widely, microbiological culture should be used more judiciously. According to the European guidelines, no urine testing should be done in low-risk women—basically adult women with recurrent cystitis or urgency symptoms or dysuria but without fever and without known diseases predisposing to urinary tract infection (UTI). "A rapid test, such as a test strip, is insensitive," Dr. Kouri says, "and culture may be sensitive but it is useless." It is usually possible to determine the organism without investigation, based on the pathogens prevailing in the community, and to prescribe antibiotics empirically. "Empiric treatment saves perhaps one-third to one-half of urine bacterial cultures," Dr. Kouri says. All other symptomatic patients belong to the high-risk group, whose specimens should be sent for bacterial culture, including identification of species and antimicrobial susceptibility testing.

Dr. Rock finds that some clinicians treat empirically based on symptoms and results of urine chemistry. "If a patient is symptomatic and the dipstick is positive for protein or leukocyte esterase or nitrite-reducing bacteria, that is a presumptive diagnosis of a bacterial UTI," he says. "If it is a first-time infection or an uncomplicated UTI, treating empirically is all right," he says. "But for repeated infections or in a patient with kidney disease you would get a culture." Dr. Rock distinguishes between office use of urinalysis and hospital or clinic use. "Office physicians are more likely to use dipstick information as a basis for treatment," he notes, "while hospital-based physicians will more likely get a microscopic examination."

In the future, Dr. Kouri foresees automated urine bacterial cultures, as well as wider adoption of automated test strip measurement. "Fully automated instruments for test strips already exist," he says. "For urine bacterial culture, there are only some big robots in research laboratories." However, he expects automation at the basic level of growth and primary typing during the next 10 or 20 years.

For now, Dr. Rock emphasizes getting the most out of urinalysis as it is currently done. He compares urinalysis to blood pressure reading. "Both are important parts of patient management," he notes, "but both need the physician to look at the result in the context of the whole patient." If he had to stress one message, he says, it would be, "Every time you order a routine urinalysis, be sure to check it closely for useful information."

William Check is a medical writer in Wilmette, Ill.