Editorial Renal Biopsies in Acute Kidney Injury: Who Are We Missing? Isaac E. Stillman,* Emerson Q. Lima, † and Emmanuel A. Burdmann † *Department of Pathology, and Renal Division, Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts; and † Division of Nephrology, Hospital de Base, Sao Jose do Rio Preto Medical School, Sao Jose do Rio Preto, Brazil Clin J Am Soc Nephrol 3: 647-648, 2008. doi: 10.2215/CJN.01110308 W hat one finds on a renal biopsy is a function of whom one chooses to biopsy. This is especially true with regards to acute kidney injury (AKI). AKI is a common entity in the hospital setting (1), even given the prob- lems of definition and terminology that have recently received attention (2), and its causes are highly dependent on the pop- ulation studied. Despite its enormous clinical impact, there have been few series of renal histology data published on the etiologies leading to AKI. Our need for more data is further complicated by the fact that the overwhelming majority of AKI cases are never biopsied. Why is this so? In many cases, the clinical context suggests the cause, often either “acute tubular necrosis” (ATN) or “prerenal,” with a reasonable degree of certainty. In others, the lack of efficient therapeutic options coupled with the risks and costs of a biopsy make it appear unwise. There are other considerations as well. In current prac- tice, most nephrologists choose to biopsy when they are not confident as to the cause of the AKI or when the renal injury has an obscure etiology. Is this appropriate? This is one context in which to approach the important contribution by Lopez-Gomez et al. appearing in this issue, detailing the latest findings from the Spanish Registry of Glo- merulonephritis, one of the largest in the world and one from which we have earlier reports (3–5). An important advantage of this registry is that the indication for biopsy was recorded, allowing identification of those biopsies performed in the set- ting of AKI, which comprised 16% of the cases. Another ad- vantage is that the findings were stratified by age. Not surpris- ingly, AKI as a biopsy indication was far more common in the elderly, although they were not the most commonly biopsied. Whether this represents reluctance on the part of the nephrolo- gists to biopsy the elderly is unclear. Given their indications for biopsy in the setting of AKI (which we assume are similar to other countries) and the fact that we are dealing with a Glomerulonephritis Registry, 90% of their patients had an “active urinary sediment,” an unusual finding in what are typically thought to be the most common causes of AKI (ATN or “prerenal”). This underscores the in- tense selection bias resulting from clinical practice inherent in any particular registry study of AKI. Presumably, clinical eval- uation led to the elimination of prerenal and postrenal causes of AKI. In essence, then, this study answers the question: “what will one find on biopsy of patients with AKI who are thought not to have ATN, but presumably glomerulonephritis?” The results deserve examination in the light of this formulation. Accordingly, one might be reassured that ATN was found in only 5% of these cases, despite its far larger presumed preva- lence in patients with AKI overall (6). Nevertheless, the devel- opment of ATN is often multifactorial, and its histologic find- ings can be subtle, raising the possibility that some cases of ATN were classified under other entities. Of course, in that case it is also likely that those entities played a role in the develop- ment of the ATN. It is also notable that the percentage of ATN was similar in all three ages, a finding that does not appear to reflect clinical experience. This report is a rare opportunity to document the causes of AKI in a selected biopsy population. However, what this study cannot answer, and what deserves asking, is the inverse ques- tion: how many patients clinically thought to have ATN actu- ally don’t, and have a different renal disease instead? We may know less than we think we do. In other words, how many patients have treatable forms of AKI that are being missed as a result of current biopsy practice? Despite the impossibility of fully answering this question, there are several studies suggest- ing significant discordance between prebiopsy and postbiopsy diagnoses in the setting of AKI. Haas et al. studied the elderly and found the clinical diagnosis to be incorrect in 34% of cases biopsied, many of them involving potentially treatable entities (7). In a more recent biopsy study, Uezono et al. found that, among elderly patients presenting with acute or rapidly pro- gressive renal injury, 71% had pauci-immune, myeloperoxi- dase–antineutrophil cytoplasmic antibody-positive, crescentic glomerulonephritis and 17% had interstitial nephritis; both groups benefited from therapeutic intervention (8). Histopatho- logic and prebiopsy clinical diagnoses differed in 15% of pa- tients, and the complication rate after biopsy was low (3%) (8). Unfortunately, the Spanish Registry did not correlate clinical and pathologic diagnosis nor report on outcome. Furthermore, Published online ahead of print. Publication date available at www.cjasn.org. Correspondence: Dr. Emmanuel A. Burdmann, Avenida Brigadeiro Faria Lima 5416, Sao Jose do Rio Preto, SP, Brazil 15090-000. Phone: 55-17-32015712; Fax: 55-17-32162227; E-mail: burdmann@famerp.br Copyright © 2008 by the American Society of Nephrology ISSN: 1555-9041/303–0647