Background: Urinalysis (UA) is considered the most important laboratory test in evaluating patients with kidney disease. Anecdotally, we have observed differences between results of UA performed by nephrologists compared with those performed by certified medical technologists or clinical laboratory scientists that could affect a clinician's diagnosis. Whether there are differences between UA performed by the clinical laboratory and that performed by a nephrologist was determined, and accuracy of diagnosis based on interpretation of the UA was compared.
Methods: Urine samples were obtained from 26 patients with acute renal failure (ARF). An aliquot of urine was sent to the clinical laboratory for UA. Nephrologist A, blinded to the patient's clinical information, performed a UA on the other aliquot of urine, generated a report, and assigned the most likely diagnosis for ARF based on UA findings. Nephrologist B, also blinded to the clinical information, reviewed nephrologist A's UA reports and assigned a diagnosis for ARF to each report. Nephrologists A and B both assigned a diagnosis (or diagnoses) for the ARF based on laboratory UA results. These 4 sets of diagnoses were compared with those assigned by the consult nephrologists.
Results: Nephrologist A correctly diagnosed the cause of ARF in 24 of 26 samples (92.3% success rate) based on his performance of the UA. Diagnoses by nephrologists A and B, based on their review of the clinical laboratory UA report, were correct in only 23.1% and 19.2% of the samples, respectively. Accuracy of diagnosis for nephrologist B improved to 69.3% when she reviewed UA reports from nephrologist A. Nephrologist A's review of urine sediment was significantly more accurate than interpretations by nephrologist A or B of clinical laboratory reports (sign test, P < 0.001). Nephrologist A reported a greater number of renal tubular epithelial (RTE) cells (P < 0.0001), granular casts (P = 0.0017), hyaline casts (P = 0.0233), RTE casts (P = 0.0008), and dysmorphic red blood cells. The laboratory noted a greater number of squamous cells (P = 0.0034).
Conclusion: A nephrologist is more likely to recognize the presence of RTE cells, granular casts, RTE casts, and dysmorphic red blood cells in urine. The laboratory may be reporting RTE cells incorrectly as squamous epithelial cells. Nephrologist-performed UA is superior to laboratory-performed UA in determining the correct diagnosis.