The Comparative Benefits of the Fractional Excretion of Urea and Sodium in Various Azotemic Oliguric StatesDiskin C.J. · Stokes T.J. · Dansby L.M. · Radcliff L. · Carter T.B.
Hypertension, Nephrology, Dialysis & Transplantation, Auburn University, Opelika, Ala., USA Corresponding Author
Background: The fractional excretion of urea (FeUrea) may result in more reliable in the determination of renal function than sodium in the presence of oliguric azotemia; however, its usefulness remains controversial, perhaps due to an evolving understanding of urea transport within the kidney. Methods: This was a prospective observational study of 100 consecutive patients referred to the nephrology service for azotemic oliguria. Multiple clinical variables were analyzed to determine variables responsible for the differences between the FeUrea and fractional excretion of sodium (FeNa) in the ability to distinguish pre-renal azotemia from intrinsic renal disease. Results: Overall, the FeUrea was more accurate (95 vs. 54%, p < 0.0001), yet both tests accurately detected the presence of intrinsic renal disease (FeNa 75%, FeUrea 85%, p = NS). The FeUrea performed significantly better (98 to 49%, p < 0.0001) in detecting pre-renal azotemia, and that advantage came exclusively in patients taking diuretics (p < 0.0001); however, 4/5 cases incorrectly detected by the FeUrea were correctly detected by the FeNa. All 4 cases had infection. Conclusion: The FeUrea appears more accurate in patients receiving diuretics; however, the FeNa may have an advantage in patients with infection.
Copyright © 2009 S. Karger AG, Basel
A common dilemma for the clinician is the need to distinguish pre-renal azotemia from intrinsic renal pathology in patients with oliguric azotemia. Although nonoliguric pre-renal azotemia can occur due to diabetes insipidus, interstitial renal disease or loss of concentrating ability from the reduced availability of urea as an intramedullary osmole, nonoliguric states are less often the cause of a pre-renal azotemia . That is because the normal renal response to an acute reduction in renal perfusion is a reduction in urine flow rate accompanied by enhanced reabsorption of sodium and urea with decreased water excretion. The fractional excretion of sodium (FeNa) developed from 17 patients with acute oliguria has proven useful for more than 3 decades [ ; however, there are many clinical situations where drugs  or disease [4,5,6] interfere with glomerular  or tubular function  which can result in an altered renal handling of sodium [1,9,10,11,12]. In an attempt to avoid these problems, clinicians have sought alternate urinary solutes to aid their diagnosis. While a fractional excretion of uric acid has been investigated , the fractional excretion of urea (FeUrea) has attracted the most attention . Although FeUrea is not affected by diuretics , recent investigators have come to diametrically opposed conclusions as to whether it offers an advantage [15,16] over the FeNa or not . Recent studies have shown that there is active transport of urea in the renal tubules. Just as diuretics can interfere with the active transport of sodium chloride and alter the FeNa, interference with active transport of urea in different clinical states may alter the FeUrea. Sepsis , gender  ], aging , protein infusion , liver disease , and certain drugs  have been suggested to interfere with the active transport of urea. A prospective study was therefore initiated to evaluate those and other variables in an effort to determine the differences between the FeUrea and FeNa and in which clinical situations are they found to be optimal diagnostic tools.
This was a prospective observational study, where 100 consecutive patients referred to the nephrology service with: (1) oliguria defined as <600 ml of urine output/24 h and (2) and elevated serum creatinine that was >1.9 mg/dl (normal 0.5–1.5 mg/dl, Jaffe reaction) were evaluated. The changes were always thought to be an abrupt alteration from normal that had been sustained for more than 24 h; however, there were 7 cases where a baseline serum creatinine was not known. Since the FeNa was originally designed only for oliguric patients , and since urea reabsorption is urine flow dependent [24,25], only oliguric patients were included. All patients with drugs or disease states that have previously reported to interfere with either (1) creatinine secretion or (2) the Jaffe reaction and produce either a falsely elevated or depressed serum creatinine [26,27] were excluded. Similarly all patients with diseases associated with spurious sodium concentration analysis performed by ion selective electrodes such as thrombocythemia, hyperlipidemia or hyperproteinemia  were excluded. All blood and urine chemistry studies were performed on Beckman Coulter DXC-800. Urea was measured by the traditional urease method. The assay for the B-type natriuretic peptide (BNP) in serum was a fluorescent immunoassay (normal 0–100 pg/ml) , which does not cross-react with any other natriuretic factor or peptide, neither does it detect Prepro BNP 22–46 nor Prepro BNP 1–21 .
Patients were analyzed by the nephrology service with complete history, physical and laboratory examinations including blood and urine samples for urea, sodium and creatinine from which FeNa and FeUrea were calculated. The original criteria described by Espinel [1 ] for the use of the FeNa were employed where <1% is considered diagnostic of pre-renal azotemia, while that of 1–3% is considered indeterminate and only >3% is considered diagnostic of acute tubular necrosis . Normally, 57% of urea is reabsorbed in the proximal tubule  and, once adjusted for inulin clearance, the FeUrea begins to drop below 50% at urine flows below 3 ml/min (4,320 ml/24 h) and appears still above 40% by the time urine flow falls below 0.41 ml/min (600 ml/24 h)  which is our definition of oliguria. Therefore, a cutoff of below 40% for FeUrea was chosen to reflect pre-renal azotemia rather than the 35% selected by previous authors [15,17] who had based their cutoff on 6 prospective patients  or diuretic induced azotemia . Since the timing of the testing in relation to the ischemic event may affect the accuracy of both indices. After the initial episode of hypotension, the kidney may still have pre-renal indices that only become those of acute renal failure after sustained ischemia . Therefore, all testing of blood and urine were made simultaneously, so should the patient’s kidney be in a transition state, both FeNa and FeUrea would be equally affected.
In only 2 cases were renal biopsies required. Clinical history, physical examination, laboratory and hemodynamic parameters such as orthostatic blood pressure, heart rate, central venous pressures and response to volume were all employed in reaching the final diagnosis by the nephrologists. Clinical course confirmed all diagnoses. By restoration of circulation to the kidneys, pre-renal azotemia returned to normal values in all patients. In patients with pre-renal azotemia due to cardiac failure, when dobutamine infusions started after the diagnosis was made, the diagnosis was not considered confirmed until the serum creatinine had remained normal for 4.5 half-lives (10–40 min) after the dobutamine was stopped since dobutamine and dopamine may spuriously lower the serum creatinine values .
Variables studied for their effect upon the FeNa and FeUrea, included: diagnosis, gender, age, and the presence of diuretics, infection, liver disease, and use of cyclosporine or total parenteral nutrition (as a surrogate for protein infusion) in addition to the urine and serum concentrations of urea, sodium and creatinine. In patients with pre-renal azotemia due to decreased cardiac output, a blood natriuretic peptide level was also analyzed to determine its utility in assessment of pre-renal status. Statistical analysis of χ2 and normality, t test for mean comparison and comparison of percentages and Pearson correlations were performed using the statistical package SPSS 12.0 (Chicago, Ill., USA).
General descriptive characteristics of the population are listed in table 1. Pre-renal azotemia occurred in 80 patients (80%) due to circulatory distress to the kidney from the following causes: cardiac failure, dehydration, sepsis, liver disease, renal artery stenosis, pulmonary embolus, and nephrogenic diabetes insipidus (table 2). The majority of patients with dehydration were taking diuretics (20/27) while other causes included nonketotic hyperglycemia (5/27) and lack of access to water (2/27).
|Table 1. General descriptive characteristics of important variables within the study population|
|Table 2. Pre-renal azotemia: causes of azotemia and effect upon the accuracy of the FeUrea and FeNa|
The FeNa and FeUrea did not assort independently but maintained a tight correlation [Pearson correlation = 0.503 (p < 0.0001)] that did not improve by removing the patients taking diuretics. While urine sodium (Una) concentrations correlated well with FeNa, no such relationship was observed with urine urea (Uu) and FeUrea. As had been previously reported , diuretics had no effect upon the relationship of Uu to FeUrea, and although removing patients with infection and liver disease improved the correlation of Uu to FeUrea, it still remained far short of significant. No patient with a Una >100 meq/dl had a FeNa that correctly detected pre-renal azotemia but no such cutoff was found for the Uu.
Overall the FeUrea was more accurate (95 vs. 54%, p < 0.0001), yet both tests accurately detected the presence of intrinsic renal disease (FeNa 75%, FeUrea 85%, p = NS) (table 3). The FeUrea performed significantly better (98 to 49%, p < 0.0001) in detecting pre-renal azotemia, and that advantage came exclusively in patients taking diuretics (p < 0.0001) (table 4). Gender, age, alkalosis, the presence of liver disease, and use of cyclosporine, aminoglycosides or total parenteral nutrition did not appear to affect the reliability of either the FeNa or FeUrea, although the number of cases in some categories were limited.
|Table 3. Intrinsic renal disease: causes and effect upon the accuracy of the FeUrea and FeNa|
|Table 4. Effect of diuretics on the accuracy of FeUrea or FeNa|
Of the 5 patients (2 pre-renal, 3 renal) incorrectly indicated by the FeUrea, 4 (80%) were correctly detected by the FeNa. All 4 had infection. Both pre-renal patients incorrectly indicated by the FeUrea had infection (2 infectious secretory diarrheas), while 2 of the 3 with intrinsic renal disease incorrectly indicated by the FeUrea had sepsis. The only patient that was incorrectly indicated by both was a patient with chronic glomerular disease. Three patients had rhabdomyolysis, one of which was mild and associated with dehydration whose oliguric azotemia resolved with fluids, while the other 2 were severe cases with prolonged oliguric renal failure requiring hemodialysis for several months. All 3 cases were correctly detected by both the FeNa and FeUrea. No cases of contrast dye nephropathy were included since in all cases, they had received acetylcysteine by protocol which had the potential of producing spuriously low creatinine values .
BNP (1,499.6 +/– 1,044.8 pg/ml) was measured in 34/39 patients with pre-renal azotemia due to a low cardiac output and was not found useful in determination of the diagnosis.
The diagnostic distinction between pre-renal and intrinsic renal disease is extremely important, since restoration of circulation to the kidneys in pre-renal azotemia will rapidly correct the problem. Although they do not assort independently, the results suggest that the FeNa and FeUrea both useful and complementary in the discrimination of these entities, depending upon the clinical situation. Diuretics clearly impaired the diagnostic ability of the FeNa, but not the FeUrea. Even in a patient with nephrogenic diabetes insipidus who had a high FeNa despite severe dehydration due to the presence of diuretics, the diagnostic capability of the FeUrea was unimpaired despite the presence of lithium that would increase urea transport, e.g. UT-A1 . Although gender, age, alkalosis, the presence of liver disease or rhabdomyolysis, and use of cyclosporine, aminoglycosides or total parenteral nutrition did not appear to affect the reliability of either the FeNa or FeUrea, sepsis and secretory diarrhea appeared to interfere with the diagnostic ability of the FeUrea.
While it has long been known that diuretics interfere with sodium chloride reabsorption in the proximal tubule, thick ascending limb and distal tubule, we investigators have only recently come to realize that urea reabsorption is also mediated by active transport. The understanding of urinary concentrating mechanisms until recently was based upon the assumption of passive transport of urea . Present understanding of urea movement across plasma membranes indicates that such movement is modulated by specialized urea transporter proteins (UT): UT-A1 and UT-A3 are now known to be active in the collecting duct, while the isoform UT-A2 is in the descending thin limb, and the isoform UT-B in the descending vasa recta . Just as diuretics can interfere with the active transport of sodium chloride and alter FeNa, drugs or disease entities that interfere with the active transport of urea will alter the FeUrea. Cytokines have been shown to downregulate urea transporters in the presence of endotoxemia from experimental infection . While FeUrea correctly distinguished between pre-renal and intrinsic renal disease in 95 (95%) patients, 4/5 (80%) of the incorrect diagnosis were in the presence of infection. Since cytokines and endotoxemia decrease FeUrea, the calculation may suggest a pre-renal etiology even in the presence of acute renal failure. Similarly, both cases of incorrect diagnoses in pre-renal patients were in patients with infectious secretory diarrhea. In the colon and small intestine, UT-B  and UT-A1 and A6  transporters have been found involved in nitrogen recycling. Despite elevated creatinines (4.2 and 3.1 mg/dl) with low FeNa (<1.0%) and urine urea concentrations (<300 mg/dl) both patients had inappropriately low blood urea nitrogen (BUN) (3 and 10 mg/dl). Since FeUrea calculation depends heavily upon BUN, which varies much more significantly than serum sodium, a low BUN resulting from loss of urea cycling in the colon may be a major pitfall for the use of FeUrea when there is infection that interferes with urea transport in the colon. Since UT-B can have divergent expression under abnormal conditions , it is theoretically possible that while sepsis might decrease FeUrea in the kidney even in acute renal failure, infectious diarrhea might result in increased colonic urea loss and, consequently, an abnormally low BUN with an elevated FeUrea even in pre-renal azotemia.
There are limitations to the conclusions that can be reached. The study population was overwhelmingly composed of patients with pre-renal azotemia (80%). Therefore, a diagnostic test that favored a pre-renal bias would therefore appear to be a more successful tool. In Spain, one series found almost the reverse ratio ; however, our incidence of pre-renal azotemia was similar to a previous report where 77% of patients studied had pre-renal azotemia . Most authorities acknowledge that pre-renal azotemia occurs twice as commonly as acute renal failure . Similarly, our study did not include nonoliguric patients as some others have. As previously noted, the FeNa was designed from data in oliguric azotemic patients and when it has been used clinically in nonoliguric patients, there have been reports of inaccuracy . Since volume conservation should be one of the hallmarks of pre-renal azotemia, nonoliguria eliminates the diagnosis of pre-renal azotemia except in a few circumstances where tubular function and concentrating ability have been altered. While a comparison of the FeNa and FeUrea might be useful in those circumstances where abnormal tubular function is expected, such a study should be done separately without mixing oliguric patients who obviously have a different physiologic situation. Finally, all patients with contrast dye nephropathy were excluded due to the use of acetylcysteine which may  (or may not ) interfere with the measurement of the serum creatinine. While this is regrettable, until the question becomes definitely settled, we could not afford to allow data that might be artificially distorted.
In summary, when employed in the proper context, the FeUrea appears to be a very useful diagnostic tool in oliguric azotemic states where there was no infection, while the FeNa remains accurate in the absence of diuretics.
Charles J. Diskin, MD
Hypertension, Nephrology, Dialysis & Transplantation
Auburn University, Suite 3, 2609 Village Professional Drive
Opelika, AL 36801 (USA)
Fax +1 334 742 0242, E-Mail HNDT512@bellsouth.net
Received: May 26, 2009
Accepted: August 17, 2009
Published online: November 3, 2009
Number of Print Pages : 6
Number of Figures : 0, Number of Tables : 4, Number of References : 46
Nephron Clinical Practice
Vol. 114, No. 2, Year 2010 (Cover Date: February 2010)
Journal Editor: El Nahas M. (Sheffield)
ISSN: 1660-2110 (Print), eISSN: 1660-2110 (Online)
For additional information: http://www.karger.com/NEC