The human eye still beats automation when it comes to analyzing urine samples for signs of kidney damage, new research presented at the National Kidney Foundation's 2014 Spring Clinical Meetings shows.
An automated urine analysis system missed signs of kidney damage that were identified by a person looking through a microscope at the sample, Dr. Natasha Sharda of the University of Arizona and her colleagues found.
Dr. Sharda and her team investigated how well each approach was able to identify granular and muddy brown casts, tiny tube-like structures made of protein secreted by the kidney. Casts in the setting of kidney damage can help identify the underlying problem such as inflammation or infection, Dr. Sharda explained, and counting them and identifying their characteristics and composition can provide important diagnostic and prognostic information.
Each of the 10 samples analyzed were from patients with acute kidney injury. A single person analyzed each urine sample, first by examining it under the microscope, and then with the automated IRIS 200 system.
The IRIS 200 system coded 70 percent of the samples as having no casts. Using manual microscopy, all of the samples were classified as having "few" casts. The samples that were automatically coded as having "many" casts were also coded as "few" by manual microscopy.
"What our research has been able to show so far is that the automated system underreported the value of granular casts in our patient cohort of acute kidney injury," Dr. Sharda said. "The automated system still has utility as a screening test, but manual microscopy should be done in all cases of abnormal kidney function, as accurate quantification of casts could have some prognostic benefit to patients."
She added: "Nowadays most hospitals use some form of an automated urinalysis system, and how often the data is actually supplemented with manual microscopy is hard to quantify. Automated systems are growing in favor as medical culture continues to place a larger emphasis towards high value cost cautious practices. Manual microscopy can take up to six minutes per sample compared to less than one minute for automated systems. Hence, labor needs and turnaround time can be dramatically reduced."
Dr. Sharda concluded: "Even though the automated urinalysis system has many benefits, it should not replace direct observation, but rather be used as augmentation."
Thomas Manley, director of scientific activities at the National Kidney Foundation, said: "Technological advances in medicine have been extremely valuable in improving outcomes for patients, but in some cases the human eye is still superior to machines. In this study manual microscopy was superior to an automated urinalysis in evaluating urine sediment properties."
He added: "It is important that clinicians are aware of the value that manual microscopic examination of urine adds in accurately detecting and evaluating kidney abnormalities. Urine testing is an important component in evaluating how the kidneys are functioning, in particular, looking for protein in the urine. The NKF strongly advocates for urine protein testing in those at risk for kidney disease, anyone with diabetes, hypertension or a family history of kidney disease."