Urine Microscopy: The Difficulty with Studying a Lost Art

The origins of urine microscopy lie in centuries past, when physicians knew little about the inner workings of the kidney and sought to understand medical problems through any accessible means. A classic example is the ancient practice of tasting urine for sweetness to detect glycosuria in patients with diabetes, first documented in the seventh century BCE by the Indian surgeon Sushruta.¹ It was not until the 17th century that pioneering physicians began looking at urine under magnification, leading to the advent of urine microscopy.^2,3 The early microscopes were rudimentary with minimal magnification power, limiting the observer to visualizing larger structures such as crystals. Eventually, cells and casts (corpuscles and cylinders) could be identified.²

Physicians soon began connecting urinary findings to clinical syndromes. Lipid droplets were seen in patients swollen from dropsy and misshapen red blood cells (RBCs) occurred in patients with Bright disease, later renamed nephrotic syndrome and glomerulonephritis, respectively.^2,4 In 1979, Birch and Fairley first demonstrated that the presence of acanthocytes differentiated glomerular from nonglomerular hematuria.⁵ Eventually, distinctive patterns of urine microscopy that correlated with specific causes of kidney disease became part of the medical lexicon: the bland urine in prerenal azotemia, muddy brown casts in acute tubular necrosis (ATN), the rarely seen white blood cells casts in interstitial nephritis, and the ever-heralded dysmorphic RBC's and RBC casts of glomerulonephritis. Over the years, many trainees have carried urine specimens in their coat pockets, waiting for rounds to finish so they could spin the urine, hoping to see Mickey Mouse ears or a Maltese cross. Setting the regulatory concerns of such practices aside, provider-performed urine microscopy (PPM) had a solid foundation in nephrology education and practice.

Although PPM thrived, it remained unclear how much the utility of urine microscopy was evidence-based and how much was common lore. With more restrictions placed on point-of-care laboratory testing, most nephrologists are unable to perform PPM and cannot rely on it to aid in clinical decision making. As of 2009, just 21% of US-based nephrology fellowship training sites performed PPM (verbal communication, Perazella, 2017). Many of these nephrologists would suggest that PPM, although nice to have, is not necessary in the modern evaluation of kidney disease. By contrast, many nephrologists who still perform urine microscopy believe it remains an essential tool.

The study by Navarro et al.⁶ sheds some light on this debate. They investigated the correlations between discrete urine microscopy findings and specific pathologic findings in a cohort of patients seen in an outpatient nephrology clinic who then went on to have a kidney biopsy. The results of a multivariate analysis were somewhat surprising and might be seen as contradictory to what has been classically taught if not taken in the proper context. Lipids were associated with mesangial abnormalities, wire loops, and hyaline deposits, but not with podocyte hypertrophy that can be seen in proteinuric states (P = 0.954). Renal tubular epithelial cells (RTECs), a nonspecific finding that classically tracks with muddy brown casts in acute tubular necrosis, were also associated with glomerular lesions such as fibrinoid necrosis and proliferative lesions.

Even more surprising was the finding that dysmorphic RBCs and RBC casts, commonly thought to be slam-dunk predictors of glomerulonephritis, did not correlate with any specific pathologic finding. As the authors point out, this may be due to the high prevalence of hematuria (54.2%) and glomerular disease (92.3%) in this cohort. Hence, two conclusions can be drawn here. There were not enough nonglomerular processes to adequately analyze this correlation and dysmorphic RBCs cannot help distinguish between subtypes of glomerular injury.

This study is enlightening because it reminds us that urine microscopy is one tool in the workup of a kidney patient and not a crystal ball. However, it does not address the question central to the debate over nephrologists performing urine microscopy: Does PPM contribute to accurately diagnosing kidney disease and aid the decision whether to perform a kidney biopsy?

Evidence in the literature supports that PPM does contribute to clinical decision making. Studies have repeatedly correlated muddy brown casts with ATN.^7,8 More broadly, Tsai et al. showed that the nephrologist correctly identified the cause of AKI 90% of the time when performing their own urine microscopy.⁹ This dropped to 23% when they relied on a laboratory-generated report. Interestingly, the accuracy was 69% when they used the report of another nephrologist's microscopy findings. This supports the notion that urine microscopy is most useful when performed by the nephrologist familiar with the case who can put the urine findings into clinical context.

This highlights a key limitation of the Navarro study. Rather than looking for patterns of injury, diagnoses, or even whether microscopy supports clinical decision making, an attempt was made to correlate individual urine microscopy findings with individual kidney pathologic findings. The study by Tsai is of particular relevance because the pathologists in this study were blinded to the patient's clinical presentation. As a result of such a granular analysis, context of the clinical case, the overall urine microscopy pattern and whether that correlates with the primary pathologic diagnosis is lost.

The major structural limitation of the Navarro study that limits its ability to answer questions about correlations between urine microscopy and kidney biopsy is the simple fact that not all patients who had urine microscopy performed went on to have a kidney biopsy. Patients with more benign urine microscopy findings are less likely to undergo a kidney biopsy and would be missing from this analysis. This severely skews the ability to use correlation as a prediction of what will be found on kidney biopsy. So rather than being a study of the association of urine microscopy findings with kidney biopsy findings, it is a retrospective look at urine microscopy findings after a kidney biopsy was performed. This difference is subtle but significant. Just as importantly, the study's narrow focus provides limited insight into the utility of urine microscopy, which is to help the clinician identify the type of kidney injury present. The generalizability of these findings is, therefore, limited, and association must not be mistaken for predictability. When performing urine microscopy in the workup of a kidney patient, the reader should not expect a kidney biopsy finding correlated to urine microscopy findings found in this study.

Although imperfect, Navarro does show us that commonly taught associations between what we see in the urine and expect on biopsy are not so lock-tight, and it is worth exploring how they may add to traditional teaching. The study highlights the broad spectrum of pathologic lesions that can be associated with lipiduria, which extends well beyond the pathologic features classically encountered in nephrotic syndrome. Closer examination of the literature (and personal experience) suggests that lipiduria can be seen in other kidney diseases, including non-nephrotic glomerulonephritis and polycystic kidney disease.^10,11 However, the lack of association with pathologic findings seen in nephrotic syndrome in this study may partially be explained by the fact that the pathologists did not have access to immunofluorescence staining results or electronic microcopy images. Relying upon podocyte hypertrophy assessed through light microscopy as the primary indicator of podocyte injury is highly subjective. Assessment of podocyte foot process effacement by electron microscopy might have demonstrated a correlation with lipiduria.

The somewhat unexpected finding that RTECs were associated with glomerular processes and not just in ATN is also supported in the literature. Tavares et al. showed that in a cohort of patients with glomerular disorders, ATN is a major driver of kidney dysfunction. ATN was seen in a majority (77%) of kidney biopsies, with elevated serum creatinine levels present in 53% of those with ATN but only 7% of those without ATN. This suggests the presence of RTECs can be expected in many cases where glomerular disease is the primary pathologic finding. ATN, even when present in a biopsy, may not be listed as a separate diagnosis when visualized alongside glomerular injury. In addition, overt ATN is inconsistently found on biopsy in patients with acute tubular injury because of its inherently dynamic state.¹¹ Thus, it is not surprising that any attempt to correlate RTECs or muddy brown casts with non-ATN processes is very difficult.

Importantly, although some of the correlations, or lack thereof, in this study were surprising and have foundation in the literature, it must be reiterated that this study does not provide evidence contradictory to classical teaching of the connection between urine findings and predicted kidney pathologic processes, such as RTECs and muddy brown casts in ATN or acanthocytes/RBC casts in glomerular processes. The limitations in this study prevent us from applying these findings more broadly.

The Navarro study is a refreshing look at the modern practice of urine microscopy and does provide some interesting insights. However, it does not call into question the classically taught correlations between urine microscopy and kidney pathology. Furthermore, it does not help answer the question about the utility of nephrologist-performed urine microscopy in the workup of kidney disease. The debate between the essential tool and the common lore camps of nephrology is likely to continue.

Disclosures

J. Simon reports the following: Consultancy: Reata—one-time consulting agreement to discuss CARDINAL study results; Patents or Royalties: UpToDate; and Advisory or Leadership Role: Alport Syndrome Foundation—Medical Advisory Committee. L. Herlitz reports the following: Consultancy: ChemoCentryx and Advisory or Leadership Role: Kidney360 editorial board.

Funding

None.

Author Contributions

J. Simon conceptualized the study and wrote the original draft; and J. Simon and L. Herlitz reviewed and edited the manuscript.