Abstract

Background. In daily clinical practice creatinine clearance is used as marker of glomerular filtration rate (GFR). As a result of the tubular secretion process endogenous creatinine clearance (ECC) overestimates glomerular filtration rate, particularly in patients with impaired renal function. It has been suggested that the tubular handling of creatinine is altered in patients with a nephrotic syndrome.

Methods. Inulin clearance (GFR) and creatinine clearance (ECC) have been simultaneously measured in a cohort of 42 patients with proteinuria and 45 healthy controls. The clearance of creatinine by tubular secretion (TScreat) can be estimated by ECC–GFR. TScreat was calculated in both groups. Regression analysis was performed to identify factors that independently influence tubular creatinine secretion.

Results. The mean age (±SD) of the patients was 41±13 years, serum albumin 26±9 g/l, median (IQR) proteinuria 4.5 (3.6–8.2) g/10 mmol creatinine, serum creatinine 103 (84–143) μmol/l, ECC 85 (69–118) ml/min/1.73 m2, and GFR 54 (36–83) ml/min/1.73 m2. Median TScreat amounted to 29 (21–36) ml/min/1.73 m2. In the healthy controls serum creatinine was 75 (70–81) μmol/l, ECC 118 (109–125) ml/min/1.73 m2, GFR 106 (102–115) ml/min/1.73 m2, and TScreat 11 (3.5–19) ml/min/1.73 m2. By regression analysis serum albumin was identified as an independent predictor of tubular creatinine secretion. We divided the patients in two subgroups based on serum albumin levels. TScreat was 24 (14–29) ml/min/1.73 m2 in patients with serum albumin levels >25.8 g/l, and 36 (28–54) ml/min/1.73 m2 in patients with serum albumin levels <25.8 g/l (P<0.01).

Conclusion. Serum albumin levels influence tubular creatinine secretion. As a result, the endogenous creatinine clearance as well as estimated GFR using a modified MDRD equation more pronouncedly overestimate glomerular filtration rate in nephrotic syndrome.

Introduction

Patients with a nephrotic syndrome are at risk to develop renal failure. Accurate assessment of renal function is important in the care of these patients. The glomerular filtration rate (GFR) can be reliably measured by using filtration markers such as inulin. Since GFR measurements with exogenous markers are generally time demanding and costly, in daily clinical practice the endogenous creatinine clearance (ECC) is often used. It is well known that in normal man the ECC overestimates the GFR as a result of the renal tubular secretion of creatinine [1]. In case of a decreased GFR the contribution of tubular secretion to total excretion of creatinine is enhanced, and as a result the overestimation of the GFR by the ECC will be more pronounced in patients with impaired renal function [2]. It has been suggested that the tubular handling of creatinine is also altered in patients with a nephrotic syndrome [2,3]. Berlyne et al. have reported four patients with a nephrotic syndrome, in whom creatinine clearance markedly overestimated GFR. However, since renal function was impaired in some of these patients, firm conclusions cannot be drawn [3]. The study of Carrie and colleagues also suggested that the creatinine clearance greatly exceeded the GFR in patients with a nephrotic syndrome [2]. However, in the latter study patients with heart failure were used as controls.

Based on the data of the Modification of Diet in Renal Disease (MDRD) study prediction equations have been developed that allow a better estimate of GFR in patients with renal diseases [4]. Recently, a simplified MDRD equation was published, which used serum creatinine as the only serum assay, and which had similar predictive ability as the original MDRD equation [5,6]. This simplified equation has already been used in large studies [7]. However, the performance of this formula in patients with hypoalbuminemia is unknown.

In the present study we have examined the relationship between ECC and GFR in patients with proteinuria. Our data indicate that tubular handling of creatinine is dependent on serum albumin levels.

Methods

For clinical studies in patients with renal diseases we regularly measured GFR using inulin clearances, according to a standardized protocol as described before [8]. All GFR measurements were approved by the hospital ethics committee, and all subjects gave their informed consent. For the present study we have analysed the data of all patients with a glomerular renal disease, and proteinuria >0.5 g/10 mmol creatinine. GFR was measured under baseline conditions. Creatinine and inulin concentrations were measured in all urine and blood samples. For comparison we have used data on ECC and GFR measured in a group of 45 healthy controls. These healthy controls were recruited from the local population and before renal measurements screened for the absence of hypertension, cardiovascular disease, renal dysfunction, and microalbuminuria. In the controls medication was not allowed except for oral anticonceptives [1].

Laboratory measurements

In blood and urine samples creatinine was determined according to a modified Jaffé method on a Hitachi 747 autoanalyser (Roche, Almere, The Netherlands). Inulin concentrations were determined in duplicate by a semi-automatic technique (centrifugal analysis, Multistat) using enzymatic degradation of inulin [9]. Albumin was measured in serum by immunonephelometry on a BNII nephelometer (Behring, Marburg, Germany) using antibodies whose specificity was checked by Ouchterlony double immunodiffusion and immunoelectrophoresis (Dako, Gloostrup, Denmark). Urinary protein was measured in 24 h urine samples using a turbidimetric method with trichloroacetic acid.

Calculations and statistics

The creatinine clearance was calculated according to the standard formula Ucr*V/Pcr in which Ucr is the creatinine concentration in the timed urine portion, V is the volume of the timed urine portion, and Pcr is the plasma concentration of creatinine measured in the same time period. The GFR was calculated with the same formula, but now using inulin concentrations instead of creatinine concentrations. The clearance of creatinine by tubular secretion (tubular clearance of creatinine; TScreat) was calculated from ECC–GFR. The simplified MDRD equation was used as follows: estimated GFR = 186 * [Plasma creatinine]−1.154 * [Age]−0.203 * [0.742 if patient is female] (plasma creatinine in mg/dl). Because all patients were Caucasian the correction factor for black people in the formula was eliminated [5]. Since serum urea was not regularly measured in our patients we were not able to use the extensive MDRD equation [4]. Total proteinuria as measured in 24 h urine samples was expressed as grams per 10 mmol creatinine to correct for errors in urine collection.

Results are given as means (±SD) or medians (interquartile range; IQR) when appropriate. For comparisons of means and medians the Student T-test or the Mann–Whitney U-test were used respectively. The Spearman correlation test was used to identify individual factors that are related to the TScreat. Next, linear regression analysis was carried out, using a forward stepwise procedure, to determine which of these individual factors independently influenced the TScreat. To allow regression analysis non-parametic parameters were transformed. TScreat showed a normal distribution after square root transformation, serum creatinine and proteinuria after log transformation. The transformed TScreat was defined as the dependent variable and the identified factors as a result of the univariate analysis were introduced as possible predicting variables. A P-value <0.05 was considered significant. All statistics were performed using SPSS software, version 11 (SPSS, IL, Chicago, USA).

Results

Data of 42 patients with proteinuria were available for analysis. The original, biopsy proven renal diseases were: membranous nephropathy (n = 23), focal glomerulosclerosis (n = 9), IgA-nephropathy (n = 4), unspecified glomerulonephritis (n = 3), membranoproliferative glomerulonephritis (n = 2), and M. Alport (n = 1). Data about medication was lacking in one patient. None of the remaining 41 patients was treated with drugs that interfere with creatinine secretion such as cimetidine or trimethoprim. No patient used steroids. Two patients were treated with NSAIDs. Twelve patients were treated with an ACE inhibitor or an AT1-receptor blocker. Clinical characteristics of the patients are summarized in Table 1. The mean age (±SD) of the 45 healthy controls (23 males, 22 females) was 28±6 years, serum albumin 44±4 g/l, the median (IQR) serum creatinine 75 (70–81) μmol/l, ECC 118 (109–125) ml/min/1.73 m2, simplified MDRD 113 (104–125) ml/min/1.73 m2, and GFR 106 (102–115) ml/min/1.73 m2.

Table 1.

Baseline characteristics of all patients and of groups of patients, ranked according to their serum albumin level

 All patients (n = 42) Serum <25.8 g/l (n = 21) Albumin >25.8 g/l (n = 21) 
Age (years) 41±13 42±14 40±13 
Sex (M:F) 35:7 17:4 18:3 
NSAID (N
ACEi/AT1B (N11/1 3/0a 8/1 
Serum albumin (g/l) 26±9 19±5b 33±6 
Serum creatinine (μmol/l) 103 (84–143) 95 (83–156) 108 (92–125) 
ECC (ml/min/1.73 m285 (69–118) 82 (63–125) 86 (74–111) 
GFR (ml/min/1.73 m254 (36–83) 43 (33–77) 63 (46–89) 
MDRD–GFR (ml/min/1.73 m268 (49–83) 68 (46–90) 67 (56–78) 
TScreat (ml/min/1.73 m229 (21–36) 36 (28–54)b 24 (14–29) 
Proteinuria (g/10 mmol creat) 4.5 (3.6–8.2) 6.0 (4.3–10.5)b 4.0 (2.7–4.9) 
 All patients (n = 42) Serum <25.8 g/l (n = 21) Albumin >25.8 g/l (n = 21) 
Age (years) 41±13 42±14 40±13 
Sex (M:F) 35:7 17:4 18:3 
NSAID (N
ACEi/AT1B (N11/1 3/0a 8/1 
Serum albumin (g/l) 26±9 19±5b 33±6 
Serum creatinine (μmol/l) 103 (84–143) 95 (83–156) 108 (92–125) 
ECC (ml/min/1.73 m285 (69–118) 82 (63–125) 86 (74–111) 
GFR (ml/min/1.73 m254 (36–83) 43 (33–77) 63 (46–89) 
MDRD–GFR (ml/min/1.73 m268 (49–83) 68 (46–90) 67 (56–78) 
TScreat (ml/min/1.73 m229 (21–36) 36 (28–54)b 24 (14–29) 
Proteinuria (g/10 mmol creat) 4.5 (3.6–8.2) 6.0 (4.3–10.5)b 4.0 (2.7–4.9) 

Data are given as means±SD or median (IQR). NSAID, number of patients treated with an NSAID; ACEi/AT1B, number of patients treated with an angiotensin converting enzyme inhibitor or angiotensin II type 1-receptor blocker; ECC, creatinine clearance; GFR, glomerular filtration rate; MDRD–GFR, GFR calculated by the Modification of Diet in Renal Disease formula, simplified version [5]; TScreat, creatinine clearance by tubular secretion.

aP = 0.062 and bP ≤ 0.01 compared to patients with serum albumin >25.8 g/l.

In the controls TScreat was 11 (3.5–19) ml/min/1.73 m2. TScreat was independent from GFR. In the patients median (IQR) TScreat was 29 (21–36) ml/min/1.73 m2 (P<0.001 vs controls) and was also not correlated with GFR (r = 0.01, P = 0.9). In univariate analysis TScreat was significantly correlated with serum albumin (r = −0.52, P<0.001). Weak correlations were found with serum creatinine (r = −0.27, P = 0.08), and proteinuria (r = 0.25; P = 0.12). Linear regression analysis was performed using the square root transformed TScreat as dependent factor and serum albumin, log transformed proteinuria, and log-transformed serum creatinine as variables. Serum albumin proved the only independent predictor. The relation between serum albumin and TScreat and between proteinuria and TScreat are depicted in Figure 1.

Fig. 1.

The clearance of creatinine by tubular secretion (TScreat) vs serum albumin (A), and vs proteinuria (B) in patients (n = 42) with proteinuria. A significant correlation was only observed between TScreat and serum albumin.

Fig. 1.

The clearance of creatinine by tubular secretion (TScreat) vs serum albumin (A), and vs proteinuria (B) in patients (n = 42) with proteinuria. A significant correlation was only observed between TScreat and serum albumin.

The use of an ACE inhibitor or AT1 blocker did not influence the results. If we restricted the analysis to patients not using an ACE inhibitor (n = 29) serum albumin proved an independent predictor of TScreat. To illustrate the potential magnitude of the effect we have analysed the data for patients divided in subgroups based on the median serum albumin level (25.8 g/l). Characteristics of the two subgroups are summarized in Table 1. TScreat was highest and thus overestimation of the GFR by the ECC was most pronounced in the subgroup of patients with the lowest serum albumin level, thus confirming the results of the linear regression analysis.

The relationship between the ECC and GFR for the two subgroups is depicted in Figure 2. It is evident that the regression lines are different. To illustrate the consequences of this difference for clinical practice, we calculated GFR for a typical patient with a measured ECC of 80 ml/min. In healthy controls an ECC of 80 ml/min/1.73 m2 represents a GFR of 60 ml/min/1.73 m2; in patients with proteinuria and a serum albumin level >25.8 g/l it represents a GFR of 57 ml/min/1.73 m2; in patients with a serum levels <25.8 g/l an ECC of 80 reflects a GFR of 42 ml/min/1.73 m2. As expected, we observed a similar difference between the patient groups when considering the relationship between calculated GFR using the simplified MDRD formula and true GFR (inulin clearance) (Figure 3).

Fig. 2.

Glomerular filtration rate (GFR) vs endogenous creatinine clearance (ECC) in patients with proteinuria and a serum albumin level <25.8 g/l (closed circles), and in patients with proteinuria and a serum albumin level >25.8 g/l (open squares). The overestimation of GFR by ECC was more pronounced in patients with low serum albumin levels.

Fig. 2.

Glomerular filtration rate (GFR) vs endogenous creatinine clearance (ECC) in patients with proteinuria and a serum albumin level <25.8 g/l (closed circles), and in patients with proteinuria and a serum albumin level >25.8 g/l (open squares). The overestimation of GFR by ECC was more pronounced in patients with low serum albumin levels.

Fig. 3.

Glomerular filtration rate (GFR) vs MDRD–GFR in patients with proteinuria and a serum albumin level <25.8 g/l (closed circles), and in patients with proteinuria and a serum albumin level >25.8 g/l (open squares). In patients with low serum albumin levels the GFR was overestimated by the MDRD–GFR.

Fig. 3.

Glomerular filtration rate (GFR) vs MDRD–GFR in patients with proteinuria and a serum albumin level <25.8 g/l (closed circles), and in patients with proteinuria and a serum albumin level >25.8 g/l (open squares). In patients with low serum albumin levels the GFR was overestimated by the MDRD–GFR.

Discussion

Our study shows that serum albumin is an independent determinant of the clearance of creatinine by tubular secretion. As a consequence overestimation of GFR by ECC is more pronounced in patients with a nephrotic syndrome. This conclusion also holds when using the simplified MDRD formula instead of ECC as predictor of GFR.

It is well known that ECC overestimates GFR, due to fact that creatinine is not only filtered but also secreted by the renal proximal tubules. Under normal circumstances tubular secretion contributes approximately 10–15% to renal creatinine clearance, the ratio of ECC/GFR amounting to 1.15 in healthy volunteers [1,10]. In patients with a decreased renal function the relative contribution of tubular secretion to renal creatinine clearance increases, which explains the widely recognized fact that ECC increasingly overestimates GFR at lower GFR [10].

Previous investigators have already pointed to the sometimes marked discrepancies between ECC and GFR in patients with a nephrotic syndrome, however, our study is the first to demonstrate the independent association between serum albumin levels and tubular creatinine handling resulting in a more pronounced overestimation of GFR. Berlyne et al. described four patients with a nephrotic syndrome and ECC/GFR ratios ranging from 1.24 to 2.37 [3]. However, the two patients with the highest ratio had markedly impaired GFR (inulin clearance). Carrie et al. studied 38 patients with a nephrotic syndrome [23]. The mean ECC/GFR ratio was 1.70±0.11. In these nephrotic patients inulin clearance was impaired. The impairment of renal function could not fully explain the increased ratio of ECC/GFR since a significantly lower ratio (1.22±0.14) was observed in patients with a comparable GFR. The latter group of “control” patients suffered form heart failure, and therefore it remained undetermined if tubular creatinine handling was altered in the nephrotic patients or in the patients with heart failure [2]. In a study in diabetic patients cimetidine, an inhibitor of creatinine transport, more pronouncedly reduced creatinine clearance in patients with macroproteinuria [11]; however, again GFR (inulin clearance) was lowest in patients with macroproteinuria. In contrast, Anderson et al. did not observe a difference in the ratio ECC/GFR between nephrotic and non-nephrotic subjects [12].

Our observations suggest that alterations in tubular creatinine handling take place as consequence of hypoalbuminaemia which can lead to major errors in the estimation of renal function in patients with a nephrotic syndrome. Our calculations indicate that in patients with a nephrotic syndrome roughly a 25% decrease of GFR may occur without any change in ECC or serum creatinine. This means that in such patients a fall in GFR will not be noticed, even by slight increases of serum creatinine.

Our study explains some discrepancies in the literature with respect to renal function parameters in patients with proteinuria. Experimental data have unequivocally shown that a reduction of albumin causes a decrease in the ultrafiltration coefficient Kf [13]. As a consequence GFR and filtration fraction are decreased when measured by precise techniques (inulin clearance or comparable methods) in patients with a nephrotic syndrome. In contrast, serum creatinine and creatinine clearance are reported as normal in most patients with minimal change nephropathy [14,15].

In recent years new equations have been developed for the estimation of GFR. Based on the MDRD data a new formula was developed, which has been validated in patients with renal failure [4]. In the MDRD formula parameters included were age, sex, race, serum creatinine, serum urea, and serum albumin. Levey et al. have subsequently published a simplified formula that included only serum creatinine as serum parameter [5]. Serum albumin and urea were excluded since these variables supposedly only contributed <1% to the observed variance of the calculations [6]. This simplified formula has been tested also in patients without renal diseases [6], and has been applied in recent studies [7]. It is clear from our study that the performance of this simplified formula is also dependent on serum albumin levels, overestimation of GFR being more prominent in patients with severe hypoalbuminaemia. Thus, a formula solely based on serum creatinine as the only serum marker should not be used in studies that include patients with severe proteinuria.

Unfortunately, serum urea levels were only available in 11 of our patients. Application of the original MDRD formula in this subgroup suggested a better performance of the original formula (data not shown). However, the paucity of data do not allow firm conclusions and larger studies are needed to validate the original MDRD formula particularly in patients with proteinuria.

It is difficult to speculate on the mechanism that may cause the altered tubular handling of creatinine in patients with low serum albumin levels. Apparently, tubular creatinine secretion is increased in patients with low serum albumin levels. Of note, serum albumin and not proteinuria was independently related to tubular creatinine handling. One mechanism could be that low serum albumin levels reflect lesser transport of albumin bound molecules that normally compete with creatinine for tubular transport.

We would like to point out an alternative explanation. We feel that our findings may be compatible with a decrease in creatinine reabsorption. In patients with a nephrotic syndrome and severe hypoalbuminemia proximal sodium reabsorption is reduced [16]. As such, creatinine reabsorption as a passive process will be influenced by changes in water and sodium reabsorption. The findings of Carrie et al. [2] with low ECC/GFR ratio in patients with heart failure also are compatible with such an idea, since in these patients creatinine reabsorption will be increased. Admittedly, creatinine reabsorption has only been demonstrated in some animal species [17]. There are data available in human studies suggesting the presence of tubular reabsorption of creatinine, particularly in patients with low rates of urine flow, however proof is lacking [18–20]. Although studies in humans are difficult, we feel that such a process cannot be excluded.

In conclusion, the present data indicate that serum albumin levels influence tubular handling of creatinine. As a result GFR is more pronouncedly overestimated by ECC in patients with a nephrotic syndrome. In patients with a nephrotic syndrome a normal serum creatinine should not be regarded as evidence of a normal GFR.

A. J. W. Branten is supported by a grant from the Dutch Science Foundation (NWO-MW 920-03-038).

Conflict of interest statement. None declared.

References

1
Vervoort G, Willems HL, Wetzels JFM. Assessment of glomerular filtration rate in healthy subjects and normo-albuminuric diabetic patients: validity of a new (MDRD) prediction equation.
Nephrol Dial Transplant
 
2002
;
17
:
1909
–1913
2
Carrie BJ, Golbetz HV, Michaels AS, Myers BD. Creatinine: an inadequate filtration marker in glomerular diseases.
Am J Med
 
1980
;
69
:
177
–182
3
Berlyne GM, Varley H, Nilwarangkur S, Hoerni M. Endogenous creatinine clearance and glomerular filtration rate.
Lancet
 
1964
;
II
:
874
–876
4
Levey AS, Bosch JP, Breyer Lewis J et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation.
Ann Intern Med
 
1999
;
130
:
461
–470
5
Levey AS, Greene T, Kusek JW, Beck GJ, MDRD Study Group. A simplified equation to predict glomerular filtration rate from serum creatinine.
J Am Soc Nephrol
 
2000
;
11
:
Abstract A0828
6
Lin J, Knight EL, Hogan M-L, Singh AK. A comparison of prediction equations for estimating glomerular filtration rate in adults without kidney disease.
J Am Soc Nephrol
 
2003
;
14
:
2573
–2580
7
Tonelli M, Moye L, Sacks FM et al. Effect of pravastatin on loss of renal function in people with moderate chronic renal insufficiency and cardiovascular disease.
J Am Soc Nephrol
 
2003
;
14
:
1605
–1613
8
Branten AJW, Wetzels JFM. Influence of albumin infusion on the urinary excretion of beta-2-microglobulin in patients with proteinuria.
Nephron
 
1999
;
81
:
329
–333
9
Degenaar CP, Frenken LA, van Hooff JP. Enzymatic method for determination of inulin.
Clin Chem
 
1987
;
33
:
1070
–1071
10
Shemesh O, Golbetz H, Kriss JP, Myers BD. Limitations of creatinine as a filtration marker in glomerulopathic patients.
Kidney Int
 
1985
;
28
:
830
–838
11
Kemperman FAW, Silberbusch J, Slaats EH et al. Estimation of the glomerular filtration rate in NIDDM patients from plasma creatinine concentration after cimetidine administration.
Diabetes Care
 
1998
;
21
:
216
–220
12
Anderson CF, Jaecks DM, Ballon HS, Palma de JR, Cutler RE. Renal handling of creatinine in nephrotic and non-nephrotic patients.
Clin Sci
 
1970
;
38
:
555
–562
13
Baylis C, Ichikawa I, Willis WT, Wilson CB, Brenner BM. Dynamics of glomerular ultrafiltration. IX. Effects of plasma protein concentration.
Am J Physiol
 
1977
;
232
:
F58
–F71
14
Korbet SM, Schwartz MM, Lewis EJ. Minimal-change glomerulopathy of adulthood.
Am J Nephrol
 
1988
;
8
:
291
–297
15
Ponticelli C, Edefonti A, Ghio L et al. Cyclosporin versus cyclophosphamide for patients with steroid-dependent and frequently relapsing idiopathic nephrotic syndrome: a multicentre randomized controlled trial.
Nephrol Dial Transplant
 
1993
;
8
:
1326
–1332
16
Grausz H, Lieberman R, Earley LE. Effect of plasma albumin on sodium reabsorption in patients with nephrotic syndrome.
Kidney Int
 
1972
;
1
:
47
–54
17
Namnun P, Insogna K, Baggish D, Hayslett JP. Evidence for bidirectional net movement of creatinine in the rat kidney.
Am J Physiol
 
1983
;
244
:
F719
–F723
18
Miller BF, Leaf A, Mamby AR, Miller Z. Validity of the endogenous creatinine clearance as a measure of glomerular filtration rate in the diseased human kidney.
J Clin Invest
 
1952
;
30
:
309
–313
19
Burcht H. On the tubular excretion of thiosulphate and creatinine under the influence of caronamide.
Scand J Clin Lab Invest
 
1949
;
1
:
270
–276
20
Perrone RD, Madias NE, Levey AS. Serum creatinine as an index of renal function: new insights into old concepts.
Clin Chem
 
1992
;
38
:
1933
–1953

Comments

0 Comments