In urinalysis, automated microscopy making the difference

“We had it in hematology, why didn’t we have it in urinalysis?” asks Darryl Hein, SH(ASCP), hematology supervisor at Children’s Hospital Central California. “It’s a no-brainer. If you want better accuracy, precision, and reproducibility, you can get them with flow cytometry,” he says.

Though the capability exists, not all urinalysis instruments are FDA approved to run body fluids, he notes. “I would prefer running body fluids on our Arkray AU-4050 Hybrid, but it is not FDA approved for that specimen type. It would be nice just to be able to run them and get an answer right away rather than having to run them on a hemocytometer.”

Within the application of flow cytometry to urinalysis, evolution is occurring on a smaller scale. Different dyes are used, separate channels are created, and more companies are getting into the game—meaning greater competition and likely even better technology.

Digital imaging, another option for automated microscopy, tends to have a lower positive predictive value compared with flow cytometry. But it makes up for that in other ways. Urinalysis is routinely performed on inpatients before admission to help determine, for billing purposes, whether the patient arrived with a UTI or acquired one during the hospital stay. “Having the pictures from digital imaging, and being able to save the pictures, is valuable from a teaching perspective and an accountability perspective,” Smith says. “I can go back and see how the staff classified cells. Before, when we were manually looking at abnormal urines under the scope, once we’d discarded the microscope slide, it was gone forever. But on the Iris, I can say, ‘He called this one a red blood cell.’
For me, the ‘look-back’ has been a game-changer.”

Automated systems don’t eliminate the need for manual microscopy, but they greatly reduce the lab’s burden. “We don’t do microscopics on 80 percent of the urine samples because either they were negative or the instrument does the microscopic for us,” says Gordon of Holy Name Medical Center. “So if we’re processing 100 urine specimens a day, only 20 of those require us to sit down and look under a microscope.”

Though some laboratories don’t perform urinalysis often enough to justify an automated instrument, other laboratories stick to manual microscopy for different reasons. “People like to put value on rare things they find with manual microscopy, but day in and day out we have lots of patients with UTIs,” Hein says. “We want to know which of those patients are going to yield a positive culture, so we need something that does a better job of counting bacteria. That’s what flow cytometry does, versus some of the other methods.”

In a small study, Hein’s laboratory looked at casts and epithelial cells from patients with known kidney disease, comparing the results from automated versus manual urine microscopy. The results were interesting, he recalls.

“The analyzer flagged for renal tubular epithelial cells—it called them ‘small round cells’ because they’re in the epithelial area of the flow cytometry, so you have to confirm that on a slide. When we looked at them on a slide, the original tech said they were just squamous. But on closer observation, there actually were renal tubular epithelial cells.” That wasn’t a surprise, he says, because the patient had known kidney disease. But if that patient had presented in the ER, the automated results would have prompted the laboratory to scrutinize the slide a bit more carefully. “The flagging feature [in the automated instrument] is a good thing,” Hein notes.

Last June, Smith’s laboratory had a similarly positive experience when it switched to the iChem Velocity urine chemistry system and the Iris iQ200 urine microscopy system. The lab had previously used a semiautomated system.

The new workflow is simple: A sample is delivered to the laboratory, poured into a tube, and placed in a rack on the instrument. The laboratorian need not return until the instrument indicates that the dipstick is negative or the microscopy is complete.

“The iQ200 uses digital imaging to look at the entire urine sample and count how many red blood cells are there, how many white blood cells are there, if there are crystals or anything pathological,” Smith says. The digital imager is remarkably accurate, she adds. “It takes really beautiful pictures of the cells. It does a pretty good job categorizing them, but the whole point is that we make sure that what the instrument is calling white blood cells are actually white blood cells, what it’s calling red blood cells are red blood cells, and so on. Once everything looks the way it’s supposed to look, we release it, and it crosses the interface, goes to the patient’s chart, and then we’re done.”

One of the best parts, Smith says, is that the laboratory’s rules are programmed into the instrument. “With their middleware solution, we were able to set up autoverification. So once results are verified, they leave the instrument. Our techs don’t have to type anything in and there are no chances for clerical errors.”

As more and more labs switch to automated urine microscopy, however, the University of Southern California’s Herrera urges them to think carefully about operator training.

“All of us know the basics of manual urinalysis,” she says. “Some people might think, ‘Oh, even a monkey can run automated urinalysis,’ but that’s not true. You still have to correlate. You still have to think about what makes sense, what’s compatible with life. The instrument is not going to tell you that. If we’re going to go with automation, we have to train people to do it properly.”

After Hershey Medical Center switched from manual to automated urine microscopy, Dr. Donaldson performed a study that revealed the “profound impacts” of automated microscopy on the clinical laboratory’s workflow and on the reproducibility of the results. “All of a sudden, instead of having variability between the amorphous stuff that might be classified as either bacteria present or absent [depending on the laboratorian], we had pretty solid repeated measures between values,” he recalls.

He has since performed other studies in hopes of securing faster, quicker, and more reliable urinalysis results that can inform the best possible clinical decisions without the need to wait for a positive urine culture.

“It’s increasingly important that we leverage our knowledge and have a delivery system that we design that gives physicians what we call ‘just-in-time information’ that’s in front of them and drives the appropriate clinical decision,” Dr. Donaldson says. “We need to make sure that if a person is on a prophylactic antibiotic therapy, it’s appropriate. And if it’s not appropriate, we need to make sure the patient gets off of it as soon as possible.”

A study of Hershey Medical Center’s urine cultures for a three-month period—a total of 13,000 cultures—assessed whether a positive or negative culture result could have been predicted on the basis of urine microscopy and chemistry findings. “We can really predict a negative result,” Dr. Donaldson says. “For a defined patient population, we can predict a urine culture result 99.9 percent of the time, which is great.”

He hopes that some day, rather than being just< a screen, urine might serve as the first portion of a reflex test to provide a quicker and more reliable tool compared with historic methods.

“There are three legs of the stool: being able to do the testing in a very controlled, very reproducible, quality-
centric way; being able to report the results efficiently; and being able to guide the physician in using those results. A comprehensive program hits all three,” Dr. Donaldson says. “In urinalysis, it’s imperative to hit all three because urine is so ubiquitous. It’s everywhere. It impacts a lot of patients in a major way, and if we don’t control for the three sets of that process, there’s a likelihood we could be reporting things that may or may not be true.”

Despite the growing popularity of automated urine microscopy, many clinicians likely remain unaware of the change in their laboratory’s urinalysis workflow. “From their side of it, it looks pretty seamless,” Smith says. “The biggest clue is that physicians are receiving the results faster because we’re not waiting for someone to enter results manually into the LIS. Instead, the result pops up on our screen when it’s ready to be validated, it’s reviewed, and then released to the physician.”

As a result, Dr. Donaldson says, some clinicians may continue to interpret urinalysis results with a grain of salt, assuming that the bacterial count is subject to a great deal of variation. “I’ve sent an e-mail out to clinicians to announce that in the last three months, there’s been a huge methodological shift at our institution. I was pretty descriptive, but if they didn’t read it then they probably don’t know that this historic variation in urinalysis results no longer exists.”

Whether or not clinicians are aware of their laboratory’s change in technology, they’re likely to take note when the internal validation studies uncover surprising results. A reference range study performed on the new instruments in Gordon’s lab, for example, revealed that certain particulates long considered abnormal, including protein aggregates known as hyaline casts, occur far more often in healthy people than previously recognized. “Normally you don’t do microscopics on normal people, and you tend to rush through slides that hardly have anything on them,” Gordon says. “Nobody had really bothered to count these things in normal people.”

“Physicians had to get used to the fact that someone can be walking around with 100 epithelials in their urine and it’s okay. Or a couple of hyaline casts—it’s okay,” Gordon says. “It changed our thinking that you should see nothing in urine.”

In the end, however, the importance of a physician’s careful evaluation cannot be overstated.

“It all comes down to the ability of a physician to have a trained eye,” Hein says. “My son is a physician in a little rural clinic, and all he has are his eyes and a dipstick. All other lab work has to be sent out. If he has a person with kidney disease, more than likely when they walk in they’re going to have edema, and they need to get on a diuretic right away, and he can see that.”

“That said, why not give them the best information we can in a hospital setting? If we could reduce patient stays with better urinalysis, I think more people would want to put more effort into getting a better sample.” [hr]

Ann Griswold is a writer in San Francisco.