Background

Fractional excretion of sodium (FE_Na) has been used in the diagnosis of acute renal failure (ARF) to distinguish between the two main causes of ARF, prerenal state and acute tubular necrosis (ATN). However, many patients with prerenal disorders receive diuretics, which decrease sodium reabsorption and thus increase FE_Na. In contrast, the fractional excretion of urea nitrogen (FE_UN) is primarily dependent on passive forces and is therefore less influenced by diuretic therapy.

Methods

To test the hypothesis that FE_UN might be more useful in evaluating ARF, we prospectively compared FE_UN with FE_Na during 102 episodes of ARF due to either prerenal azotemia or ATN.

Results

Patients were divided into three groups: those with prerenal azotemia (N = 50), those with prerenal azotemia treated with diuretics (N = 27), and those with ATN (N = 25). FE_Na was low only in the patients with untreated plain prerenal azotemia while it was high in both the prerenal with diuretics and the ATN groups. FE_UN was essentially identical in the two pre-renal groups (27.9 2.4% vs. 24.5 2.3%), and very different from the FE_UN found in ATN (58.6 3.6%, P < 0.0001). While 92% of the patients with prerenal azotemia had a FE_Na <1%, only 48% of those patients with prerenal and diuretic therapy had such a low FE_Na. By contrast 89% of this latter group had a FE_UN <35%.

Conclusions

Low FE_UN (35%) was found to be a more sensitive and specific index than FE_Na in differentiating between ARF due to prerenal azotemia and that due to ATN, especially if diuretics have been administered.

METHODS

This was a prospective, Institutional Review Board approved study that took place at Nassau County Medical Center, a large (612 beds) public hospital. One hundred two consecutive adult patients referred to the Nephrology service (mostly from the Intensive Care Units) for evaluation of acute renal failure, comprised the subjects of this study. The major reason for consultation was a rapidly increasing blood urea nitrogen (BUN) and creatinine (BUN>30 mg/dL and creatinine>1.5 mg/dL) with or without oliguria. Urine output measurements were available for most patients, in particular all the ATN patients who were typically found in one of the ICUs, all of whom had Foley catheters. In all cases, the diagnosis was made by the consulting attending nephrologist, based on the medical history, physical examination, and evaluation of urine sediment morphology and composition, as suggested by standard textbooks¹, and is shown in Table 1. The members of the renal service used the usual urinary indices as described in Table 1. Urinary urea nitrogen was not available at the time of consultation, but was measured later in batches of urine specimens from 5 to 10 patients. Twenty-seven of the patients had received diuretic therapy (furosemide or thiazides, up to the day of consultation). Patients with acute interstitial nephritis, acute glomerulonephritis, and obstructive nephropathy were not included in the present study as there were a small number of patients in each category (3, 1, and 1, respectively) that would not allow for drawing any meaningful conclusions.

Table 1 - Criteria for diagnosis of acute renal failure and to allow to differentiate prerenal azotemia and acute tubular necrosis (ATN).

Full table

Prerenal azotemia was the result of circulatory failure secondary to various causes such as sepsis, gastrointestinal bleeding, hepatic, respiratory or cardiac failure. The diagnosis of prerenal azotemia was established in circulatory failure when improvement of heart function, cessation of diuretic therapy, or treatment of shock effected a prompt increase in urinary output and creatinine clearance. Of the 50 cases of prerenal azotemia, 16 had sepsis, 10 congestive heart failure, 7 hepatic failure and 7 gastrointestinal bleeding. The remaining patients had miscellaneous or less definite diagnoses. Twenty-four were men and 26 were women, with an average age of 49 2 years (mean SEM). The demographics were actually quite similar in all subgroups where the representation of men to women were 14 to 13 in the group of prerenal on diuretics and 13 to 12 in the ATN group. The mean ages were 51 3 and 47 3 in these two groups, respectively.

As expected, the majority of the prerenal cases receiving diuretics were subjects with congestive heart failure. Acute tubular necrosis (ATN) resulted from nephrotoxic drugs (aminoglycoside antibiotics), myoglobinuria, sustained circulatory failure, or massive surgical volume losses unresponsive to replacement. Thirteen patients had protracted sepsis, although the exact role of coexistent factors such as major surgical procedures (8 of these) or use of nephrotoxins (5 were on aminoglycoside therapy) does not allow for a clear-cut identification of the underlying etiology. Myoglobinuria caused six cases of ATN (3 due to cocaine, 1 due to alcohol, 1 due to phosphorus depletion and 1 due to trauma). The remainder of the cases had a mixed and often less well identified etiology.

The indices evaluated at the time of consultation were urinary sodium (U_Na), urinary sodium/potassium ratio (U_Na/K; a reflection of prerenal conditions with associated hyperaldosteronism), urinary to plasma creatinine ratio (U/P_Cr), and fractional excretion of sodium (FE_Na). Fractional excretion of urea (FE_UN) was determined later. The FE_Na was calculated as [(urine sodium/ plasma sodium)/(urine creatinine/plasma creatinine)] 100. The FE_UN was calculated as: [(urine urea nitrogen/ blood urea nitrogen)/(urine creatinine/plasma creatinine)] 100. In well-hydrated individuals, the FE_UN is 50 to 65%¹⁹. In the presence of oliguria and prerenal azotemia, FE_UN less than or equal to 35% with progressive decline has been found as urine volume decreases to about 0.5 mL/min or less¹⁹. Indeed, as urine volume decreases to oliguric levels (<0.35 mL/min or 500 mL/day) FE_UN decreases proportionally to urine flow²⁰. For that reason, we selected a FE_UN value of 35% to signify prerenal azotemia. On the other hand, in intrinsic renal failure the clearance of urea decreases to an equal degree or less than the clearance of creatinine. Thus, an FE_UN more than 50% indicates acute tubular necrosis. Blood chemistries were performed on a Beckman automated analyzer for serum electrolytes, creatinine and urea (Beckman Instruments, Fullerton, CA, USA). Urine electrolytes were determined by flame photometry. Urine urea nitrogen and creatinine were determined spectrophotometrically on spot urine samples collected at the bedside. In one part of the study the measured FE_UN (as described earlier in this article) was compared with the expected FE_UN from the study of Dole, where FE_UN = 0.57 e^-(0.36/V)19. In a few prerenal cases where accurate (urine flow in mL/min) was difficult to assess, the calculated from U/P_Cr was used with = 100/(U/P_Cr) as described previously by Homer Smith²¹.

Statistical analysis was performed using the appropriate statistical method²². One way analysis of variance (ANOVA) was used when more than two groups were evaluated. Comparison among the groups was tested using the Bonferroni-corrected t test. In paired experiments between two groups, the paired Student t test was used. Linear regression was calculated by using the least squares method and the Pearson product moment analysis. All tests were performed using Sigma Stat®, a statistical software program of the Jandel Corporation (Corte Madera, CA, USA). In tests presented in nominal (categorical) form, we determined the sensitivity, the specificity, the positive predictive value, and the negative predictive value. These tests also were compared by using the receiver operated characteristic (ROC) curves²².

RESULTS

The patients were divided into three groups. The first group was composed of 50 patients with prerenal failure, the second group had 27 patients with prerenal azotemia but also given diuretics (furosemide), and the third group had 25 patients with established ATN. The indices evaluated in ARF are shown in Table 2. It is clear that while FE_Na was characteristically low (<1%) in the prerenal group, it was substantially higher than 1% in both the prerenal group given diuretics and the ATN group. In contrast, the increased BUN/creatinine ratio, the urine-to-plasma creatinine ratio (U/P_Cr) and, more impressively, the FE_UN, were consistently different in both prerenal groups from those in the ATN group Figure 1. Indeed, the FE_UN was essentially identical in the two prerenal groups (27.9 2.4% vs. 24.5 2.3%), significantly different from the FE_UN found in ATN (58.6 3.6%, P < 0.0001).

Figure 1.

Levels of blood urea nitrogen/creatinine (BUN/Cr; A), fractional excretion of sodium (FE_Na; B), fractional excretion of urea nitrogen (FE_UN; C), and urinary plasma to creatinine ratio (U/P_Cr; D) are schematically represented for each category of acute renal failure. BUN/Cr and U/P_Cr are ratios and have no units. FE_Na and FE_UN are expressed as %. Please note that FE_UN is similar among prerenal patients, irrespective of their use or non-use of diuretics. Note a rather substantial variability with other indices.

Full figure and legend (64K)

Table 2 - Evaluation of the significance of the different indices in the different forms of acute renal failure.

Full table

Table 3 and Figure 2 show the importance of the indices in patients in whom a given test was positive. It appears that FE_Na, FE_UN, and U/P_Cr are the indices that discriminate best among the different types of acute renal failure. While 92% of the patients with prerenal azotemia had a FE_Na <1%, only 48% of patients who were prerenal and on diuretic therapy had a low FE_Na. By contrast, 89% of this latter group had a FE_UN <35%. If one combines all the prerenal cases, FE_UN had the best sensitivity and specificity (90 and 96%, respectively), and the best positive and negative predictive value (99 and 75%, respectively) of all the indices examined in this study.

Figure 2.

Sensitivity of the three major indices for patients with prerenal situations with or without diuretic use, and patients with ATN. The units shown on the vertical axis represent %. Please note that while all three indices are equally effective in patients with prerenal situations who did not use diuretic therapy, only FE_UN retains its high sensitivity even in prerenal patients who use diuretics. Symbols are: () FE_Na; () FE_UN; () U/P_Cr.

Full figure and legend (37K)

Table 3 - Number of patients with positive urinary indices in each ARF group.

Full table

Table 4 evaluates the positivity of different levels of the three most important indices in patients with acute renal failure. Again, when examining the combined prerenal group compared with that of ATN, FE_UN was the best discriminatory index. For instance, a FE_Na of less than 1% was present in 77% of prerenal patients and in about 4% of patients with ATN, while FE_UN <35% was found in 90% of prerenal patients and again in 4% of patients with ATN. U/P_Cr appears to be an intermediate differentiator. When a ratio higher than 15 was used as the cutoff value, 90% of patients with prerenal azotemia were positive, with only a small number of patients with ATN having such a level (8%). Figures 3 and 4 show the ROC curves for these data.

Figure 3.

Sensitivity and false positivity of different cut of levels are plotted for FE_Na () and U/P_Cr (). The levels used are the same that are shown for the two parameters in Table 3. Sensitivity was defined by the % of patients who were positive for prerenal (either with or without diuretics) at a given level of each test. False positive was defined as the % of patients with ATN who had pre-renal values for any level.

Full figure and legend (19K)

Figure 4.

Sensitivity and false positivity of different cut of levels are plotted for FE_UN () and FE_Na (). The levels used are the same that are shown for the two parameters in Table 3. Sensitivity was defined as the % of patients who were positive for prerenal (either with or without diuretics) at a given level. False positive was defined as the % of patients with ATN who had prerenal values for any level. While the area under FE_UN is 97.2%, that under FE_Na is significantly less (88.9%).

Full figure and legend (22K)

Table 4 - Number of patients at different levels of the three most significant indices.

Full table

We evaluated the equations that describe the relationship between the FE_UN and U/P_Cr in our subjects. This was based on previous findings regarding such a relationship (see discussion below for details). We found major differences among the different groups. In prerenal azotemia, FE_UN = 27.9–0.029 (U/P_cr), r = 0.28, P = 0.015, and was similar for the two groups of prerenal patients (prerenal alone or prerenal in the presence of diuretic therapy). In patients with ATN a very different relationship was noted: FE_UN = 70.3–1.26 (U/P_Cr), r = 0.439, P = 0.028, consistent with very different mechanisms of the generation of such a relationship.

Finally, the predicted FE_UN was examined using Dole's equation as described in the Methods section. In the prerenal patients, the expected FE_UN was 25 4% and 22 4% for the prerenal group without and with diuretic use, respectively. These values were not statistically different from the ones actually found (27 2% and 24 3%, respectively), suggesting that indeed the mechanism for the decreased volume (decreased renal perfusion and increased water reabsorption) also was responsible for the decreased excretion of urea. In contrast, in the cases of ATN the calculated FE_UN was very different from the one actually determined (18 3% vs. 58 6%, P < 0.0001), suggesting that the oliguria found in this condition is not related to reabsorption of water and urea from functional tubules, but rather due to minimal glomerular filtration through injured nephrons.

DISCUSSION

When urine excretion decreases as a result of water reabsorption, the level of urinary creatinine increases in inverse proportion to the volume of urine output. Thus, the U/P creatinine ratio logically identifies whether oliguria is the result of increased water reabsorption, as in the case of prerenal situations (U/P_Cr>20) or is due to the loss of nephron function (U/P_Cr <20)²¹. As is often appreciated, sodium reabsorption also increases in prerenal states both as a result of an increase in aldosterone as well as the increased sodium and water reabsorption in the proximal tubule because of the increased filtration fraction found in such cases. A composite index has been developed and used clinically that incorporates both U/P_Cr and U_Na, the FE_Na. Following the initial report of Espinel⁴, which showed that FE_Na discriminates best between the two most commonly encountered forms of ARF (prerenal and ATN), several subsequent reports substantiated this claim^1,2,3,4. Currently it has become the dominant discriminatory index used for that purpose. On the other hand, coexistence of a prerenal state and natriuretic agents (diuretics) negate the effectiveness of this index¹. Diuretics represent the mainstay of treatment for several prerenal states and may themselves produce volume depletion due to their natriuretic effect (even if given to initially euvolemic patients). Furthermore, these agents are used commonly in the hospital and even more in the intensive care unit, thus significantly decreasing the utility of FE_Na in the hospital setting. Hence, FE_Na does not give a clear-cut distinction in such cases. In addition to the use of diuretics, other situations are known to lead to both prerenal azotemia and increased FE_Na. Better understood among them is the case of volume depletion due to vomiting or naso-gastric suction associated with increased FE_Na due to bicarbonaturia¹⁸. On the other hand, a number of clear-cut forms of ATN are associated with decreased FE_Na, for example: myoglobinuria^6,7, sepsis⁸, contrast nephropathy^11,12,13, obstructive uropathy¹⁵, and non-oliguric ATN^2,3,5. Low FE_Na also is found in other forms of acute intrinsic renal failure including acute glomerulonephritis^2,3, urinary tract obstruction², and renal allograft rejection². To emphasize the frequency with which patients with intrinsic acute renal failure have FE_Na <1%, suffice it to say that in a five year-period (1980 to 1985) more than 100 cases of documented ATN with FE_Na <1% were found in the published literature alone¹⁷.

The relative clearance of urea denoted by the FE_UN is affected by decreased renal perfusion as seen in prerenal states. As early as 1921, Austin, Stillman and Van Slyke showed that the clearance of urea in man decreases substantially in dehydration somewhat proportionally to urine output²³. In 1938, two seminal studies established that FE_UN relates inversely to U/P creatinine²⁴ and U/P inulin^25,26, and therefore proportionally to urine volume. This was the classical teaching in renal physiology during that era^27,28. More recently, Goldstein, Lenz and Levitt evaluated the relationship of urine volume to urea excretion in humans²⁹. Using proximal and distal diuretic agents, they found that the percent of filtered urea excreted increased with proximal (mannitol and acetazolamide) but not with distal agents (mercaptomerin and ethacrynic acid). This suggests that urea is not reabsorbed significantly in the distal nephron. Given that most of the diuretics used clinically work at distal sites, FE_UN should not be affected by their use. This is in contrast to FE_Na, which is increased by all forms of diuretics. Only in patients with established chronic renal failure does the use of diuretics and the ensuing volume depletion also operate in distal sites³⁰, a finding with limited significance in the differential diagnosis of ARF.

In 1992, Kaplan and Kohn reported six prerenal patients with a discrepancy of FE_UN and FE_Na in the presence of diuretics³¹. In addition to the small number, most of those patients already had established renal failure, congestive heart failure and various treatments with angiotensin-converting enzyme inhibitors. The authors, in order to better understand the situation, retrospectively evaluated the charts of 87 patients with prerenal azotemia. They noted that the patients that had both low FE_Na and FE_UN had the usual type of prerenal azotemia. The group with low FE_UN but high FE_Na appeared to have prerenal azotemia but had been treated with a diuretic (39 out of 40). The authors concluded that FE_UN may be a better index to evaluate possible prerenal status in patients who receive diuretic agents.

We prospectively studied three groups of patients: patients with prerenal azotemia, patients with prerenal conditions receiving diuretics, and patients with acute tubular necrosis (ATN). We found that FE_Na failed to discriminate between prerenal patients given diuretics and those with ATN. FE_UN, on the other hand, was excellent in discriminating among all cases of prerenal azotemia and ATN irrespective of the use of diuretics. This has a potentially major practical application, given the frequency of diuretic use in the hospital setting. For instance, among our 77 unselected prerenal cases, 27 (35%) had been given diuretics.

Our study found that U/P_Cr is higher in prerenal cases, as has been shown before and as expected from the underlying physiology. Indeed, it appears that U/P_Cr is at least as good as FE_Na in separating prerenal azotemia from ATN. As was previously reported^{24,25,26,27,28}, there was a close inverse relationship between U/P_Cr and FE_UN in the prerenal group. This relationship was very different in the group with ATN. Furthermore, the prerenal patients appear to follow the prediction of Dole¹⁹ in reference to expected FE_UN based on urine output. In short, these patients behaved the same way as a normal individual with decreased fluid intake and thus decreased urine output. In contrast, urine volume could not account for the FE_UN found in the ATN group.

We found that FE_UN has a high sensitivity (85%) a high specificity (92%) and more importantly a high positive predictive value (98%). The latter suggests that if a FE_UN of <35% is found, 98% of such cases will have prerenal azotemia. The relative importance of FE_UN to FE_Na would have been more impressive if our ATN patients had had some of the characteristics described in the literature. For example, low FE_Na is often found in patients with myoglobinuric ATN^6,7; however, in our study none of the six patients with well-documented myoglobinuria (history and urine analysis positive for myoglobin) had an FE_Na value less than 1. If the value had been greater, then the false-positive cases of low FE_Na in identifying pre-renal azotemia would have been even higher.

Fractional excretion of urea nitrogen to a great extent relates inversely to the proximal reabsorption of water. Proximal reabsorption of water (accompanied by urea) increases when renal perfusion decreases and filtration fraction increases. Thus, urea reabsorption leads to a decrease in FE_UN and an increase in the BUN/creatinine ratio. The use of indices that reflect proximal tubular function in acute renal failure have been suggested before. A recent textbook of nephrology suggests two such indices, the clearance of lithium and the clearance of uric acid³². It is apparent that the use of lithium clearance is quite impractical, given the limited availability of this determination in many laboratories and the need to administer lithium to such patients. Uric acid clearance does not appear to have any advantage over the use of urea, due to its rather complex renal metabolism.

It is certainly interesting, that if we modify the formula for FE_UN = (U/P)UN/(U/P)Cr, it becomes FE_UN = U (UN/Cr)/(BUN/Cr), or even, BUN/Cr = U (UN/Cr)/FE_UN. This suggests that as proximal urea reabsorption increases and FE_UN decreases in prerenal states, the BUN/Cr ratio increases, as is the common teaching and is our own experience in the present study Table 2. Another cause of increased BUN/Cr is increased urea production (and thus excretion), as in cases of hypercatabolism or consumption of high protein diets. This could be easily differentiated from prerenal azotemia, as in high urea production we have a high urine UN/Cr, much higher than the usual ratio of 5 to 10. The determination of urine urea nitrogen and FE_UN may be helpful in a few other conditions. For instance, the presence of a BUN/Cr ratio of 10 in the presence of a low U_UN/U_Cr (<5) suggests the existence of prerenal azotemia in a starving individual.

There are situations, however, where FE_UN may fail to indicate the ongoing prerenal status. Therefore, such possibilities should be kept in mind when FE_UN is used. For instance, in osmotic diuresis the proximal tubular absorption of salt and water is impaired and thus increased FE_UN is expected despite renal hypoperfusion. One of us (C.P.C.) has looked into this while working on another study (abstract; Vafadouste R et al, XV International Congress of Nephrology, Buenos Aires, Argentina, p 1999, 2001). In ten patients with osmotic diuresis there was a mild increase of BUN (23 3.3 mg/dL) with normal serum creatinine (0.9 0.1 mg/dL), increased BUN/Cr to 27 3, increased FE_UN 49 8%, high normal FE_Na 0.7 0.3% and high U/P_Cr of 41 10. The same phenomenon is expected to happen when osmotic diuretics such as acetazolamide and mannitol are used. Indeed, this was our finding during that study. Similar findings also were reported before²⁹. A similar picture emerges in patients given a high protein diet or having excessive catabolism. There, the enormous production of urea acts as the osmotic diuretic agent. On the other hand, FE_UN could be helpful in cases with prerenal azotemia in combination with a distal defect in sodium reabsorption. We recently encountered a young male with AIDS and Addison's disease. He had apparent volume depletion with dramatic changes in blood pressure and pulse rate with even the mildest tilting maneuvers. His BUN/Cr was high (84/2.3), and he had a high FE_Na (>2%), a FE_UN of 35% and a U/P_Cr of 27. This patient responded promptly to vigorous saline infusion, indicating that his major problem was that of volume depletion in the presence of impaired distal tubular Na⁺ reabsorption.

Although FE_UN proved to be the best index under the clinical conditions encountered in our study, one must remember that none of the indices offers a 100% discriminatory ability. The three major indices allow evaluation of different sites of action, and differ in the technology needed for their determination, thus making it unlikely to have simultaneous laboratory errors in all of them. FE_UN allows insight to proximal peritubular forces, FE_Na mostly reflects distal forces (mostly aldosterone-dependent sites), and U/P_Cr indicates water reabsorption through the length of the nephron. Indices are tools and do not obviate the importance of history, physical examination, and direct urinalysis in making a diagnosis of acute renal failure.

References

REFERENCES

1. Brady, HR, Brenner, BM, Lieberthal, W: Acute renal failure, in Brenner & Rector's The Kidney (1996, 5^th ed), edited by Brenner BM, Philadelphia, W.B. Saunders, pp 1200–1252
2. Zarich, S, Fang, LS, Diamond, JR: Fractional excretion of sodium. Exceptions to its diagnostic value. Arch Int Med 1985 145: 108–112,  | ISI | ChemPort |
3. Miller, TR, Anderson, RJ, Linas, SL, et al: Urinary diagnostic indices in acute renal failure. Ann Intern Med 1978 89: 47–50,  | PubMed | ISI | ChemPort |
4. Espinel, CH: The FENa test: Use in the differential diagnosis of acute renal failure. JAMA 1976 236: 579–581,  | Article | PubMed | ISI | ChemPort |
5. Danovitch, G, Carvounis, C, Weinstein, E, Levenson, S: Non-oliguric acute renal failure. Isr J Med Sci 1979 15: 5–8,  | PubMed | ISI | ChemPort |
6. Steiner, RW: Low fractional excretion of sodium in myoglobinuric acute renal failure. Arch Intern Med 1982 142: 1216–1217,  | ISI |
7. Corwin, HL, Schreiber, MJ, Fang, ST: Low fractional excretion of sodium. Occurrence with hemoglobinuric and myoglobinuric acute renal failure. Arch Intern Med 1984 144: 981–982,  | Article | PubMed | ISI | ChemPort |
8. Vaz, AJ: Low fractional excretion of urine sodium in acute renal failure due to sepsis. Arch Intern Med 1983 143: 738–739,  | Article | PubMed | ISI | ChemPort |
9. Pru, C, Kjellstrand, CM: On the clinical usefulness of the FE_Na test in acute renal failure: A critical analysis. Proc Dial Transplant Forum 1980 10: 240–246,  | ChemPort |
10. Pru, C, Kjellstrand, CM: The FE_Na test is of no prognostic value in acute renal failure. Nephron 1984 36: 20–23,  | PubMed | ISI | ChemPort |
11. Fang, LST, Sirota, RA, Ebert, TH, et al: Low fractional excretion of sodium with contrast media-induced acute renal failure. Arch Intern Med 1980 140: 531–533,  | Article | PubMed | ISI | ChemPort |
12. Carvallo, A, Raboski, TA, Argy, WP, et al: Acute renal failure following drip infusion pyelography. Am J Med 1978 65: 38–45,  | Article | PubMed | ISI | ChemPort |
13. VanZee, B, Hoy, W, Talley, TE, et al: Renal injury associated with intravenous pyelography in non-diabetic and diabetic patients. Ann Intern Med 1978 89: 51–54,  | PubMed | ISI | ChemPort |
14. Oken, DE: On the differential diagnosis of acute renal failure. Am J Med 1981 71: 916–920,  | Article | PubMed | ISI | ChemPort |
15. Hoffman, LM, Suki, WN: Obstructive uropathy mimicking volume depletion. JAMA 1976 236: 2096–2097,  | Article | PubMed | ISI | ChemPort |
16. Steiner, RW: Interpreting the fractional excretion of sodium. Am J Med 1984 77: 699–702,  | Article | PubMed | ISI | ChemPort |
17. Pru, C, Kjellstrand, CM: Urinary indices and chemistries in the differential diagnosis of prerenal failure and acute tubular necrosis. Semin Nephrol 1985 5: 224–233,  | PubMed | ISI | ChemPort |
18. Nanji, AJ: Increased fractional excretion of sodium in prerenal azotemia: Need for careful interpretation. Clin Chem 1981 27: 1314–1315,  | PubMed | ISI | ChemPort |
19. Dole, VP: Back diffusion of urea in the mammalian kidney. Am J Physiol 1943 139: 504–519,  | ChemPort |
20. Chesley, LC: Urea excretion at low urine volumes. The calculation of 'minimal' urea clearances. J Clin Invest 1938 17: 119–138,  | ChemPort |
21. Smith, HW: Principles of Renal Physiology. 1956, New York, Oxford University Press, pp 79–81
22. Carvounis, CP: Medical Statistics. 2000 London, Parthenon Publishing Group, Inc.,
23. Austin, JH, Stillman, E, Van Slyke, DD: Factors governing the excretion of urea. J Biol Chem 1921 46: 91–102,  | ChemPort |
24. Chasis, H, Smith, HW: The excretion of urea in normal man and in subjects with glomerulonephritis. J Clin Invest 1938 17: 347–358,  | ChemPort |
25. Bull, GM, Joekes, AM, Lowe, KG: Renal function studies in acute tubular necrosis. Clin Sci 1950 9: 379–404,  | ISI |
26. Shannon, JA: The renal reabsorption and excretion of urea under conditions of extreme diuresis. Am J Physiol 1938 123: 182–196,
27. Smith, HW: The Kidney. Structure and function in health and disease. 1951, New York, Oxford University Press, pp 63–80
28. Pitts, RF: Kidney and Body Fluids (1968, 2^nd ed). Chicago, Year Book Medical Publishers Inc., pp 88–93
29. Goldstein, MH, Lenz, PR, Levitt, MF: Effect of urine flow rate on urea reabsorption in man. Urea as a 'tubular marker'. J Appl Physiol 1969 26: 594–599,  | PubMed | ISI | ChemPort |
30. Dal Canton, A, Fuiano, G, Conte, G, et al: Mechanism of increased plasma urea after diuretic therapy in uremic patients. Clin Sci 1985 68: 255–261,  | PubMed | ChemPort |
31. Kaplan, AA, Kohn, OF: Fractional excretion of urea as a guide to renal dysfunction. Am J Nephrol 1992 12: 49–54,  | PubMed | ISI | ChemPort |
32. Dwinnell, BG, Anderson, RJ: Diagnostic evaluation of the patient with acute renal failure, in Atlas of Diseases of Kidney, 1999, edited by Schrier RW, Philadelphia, Current Medicine Inc., pp 12.1–12.12

Acknowledgments

This study was presented in part in the 33^rd Annual Scientific Meetings of the American Society of Nephrology in Toronto, Canada, on October 13, 2000. The authors acknowledge their gratitude to Drs. Donald A. Feinfeld, Thomas Manis, and Laurie Ward for their critical evaluation of this manuscript and their many suggestions during the study