Skip Navigation


NDT Advance Access originally published online on August 25, 2006
Nephrology Dialysis Transplantation 2006 21(12):3488-3494; doi:10.1093/ndt/gfl430
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
21/12/3488    most recent
gfl430v2
gfl430v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Related articles in NDT
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Macdonald, J. H.
Right arrow Articles by Lemmey, A. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Macdonald, J. H.
Right arrow Articles by Lemmey, A. B.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2006]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

The relationship between estimated glomerular filtration rate, demographic and anthropometric variables is mediated by muscle mass in non-diabetic patients with chronic kidney disease

Jamie Hugo Macdonald1, Samuele Maria Marcora1, Mick John Kumwenda2, Mahdi Jibani3, Gareth Roberts3, Ruth Glover2, Jeffrey Barron4 and Andrew Bruce Lemmey1

1School of Sport, Health and Exercise Sciences, University of Wales, Bangor, George Building, Bangor, Gwynedd LL57 2PZ, 2Renal Unit, Ysbyty Glan Clwyd, Rhyl, Denbighshire LL18 5UJ, 3Renal Unit, Ysbyty Gwynedd Penrhosgarnedd, Bangor, Gwynedd LL57 2PW and 4Department of Chemical Pathology and Metabolism, St Helier Hospital, Carshalton, Surrey SM5 1AA, UK

Correspondence and offprint requests to: Jamie Hugo Macdonald School of Sport, Health and Exercise Sciences, University of Wales, Bangor, George Building, Bangor, Gwynedd LL57 2PZ, UK. Email: j.h.macdonald{at}bangor.ac.uk



   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Acknowledgements
 References
 
Background. In this study (the first of two related papers), we report whether the relationship between the demographic and anthropometric variables (DA, i.e. age, gender, height and weight) employed in current creatinin (Cr)-based glomerular filtration rate (GFR) estimation equations and actual GFR is mediated by muscle mass.

Methods. We studied 77 patients (mean age ± SD, 65.1 ± 11.9 years) with chronic kidney disease (mean GFR 45.7 ± 28.6 ml/min/1.73 m2). Actual GFR was measured by the renal clearance of inulin (GFRinu). Appendicular lean mass (ALM) and its index (ALMI) by dual energy X-ray absorptiometry provided markers of muscle mass. Multiple regression analyses identified variables explaining variance in (i) GFR, (ii) ALM and (iii) Cr.

Results. (i) The DA variables used in the abbreviated modification of diet in renal disease (MDRD) equation accounted for only 59.6% (P < 0.001) of the variance in GFRinu, whilst adding ALMI explained an additional 10.4% variance (P < 0.001). If ALMI was entered first, the relationship between DA variables and GFRinu was reduced (for weight) or completely abolished (for age, gender and height). (ii) After inputting all the commonly used DA variables, 17.2% of the variance in ALM was unexplained. (iii) All the DA variables explained only 60.6% (P < 0.001) of the variance in Cr, whilst adding ALM explained an additional 4.2% variance (P < 0.005).

Conclusions. Muscle mass explained more variance in GFRinu than MDRD DA variables and mediated the relationship between GFRinu and DA variables. Furthermore, DA variables failed to account for individual differences in muscle mass or Cr. Consequently, there is a need to validate simpler, clinically obtainable measures of muscle mass and determine whether these measures will improve GFR estimation.

Keywords: chronic kidney disease; creatinine; DXA; glomerular filtration rate; MDRD; muscle mass



   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Acknowledgements
 References
 
Accurate measures of glomerular filtration rate (GFR), which provide a useful index of disease severity, are required to help decide when to start dialysis and to estimate correct drug dosage. Yet gold standard measures of GFR remain cumbersome and expensive. Consequently, clinicians and researchers estimate GFR using one of a variety of available prediction equations. These prediction equations include serum creatinine (Cr) because Cr is excreted principally by the kidneys and Cr can be determined cheaply and quickly in clinical practice [1]. Previous researchers [2–4] have identified that additional variables such as age, gender, ethnicity, height and weight also explain a significant proportion of the variance in kidney function. It is generally accepted that these variables directly affect kidney function [5–7] and hence these variables, with Cr, are used to predict GFR. However, these variables are also predictors of skeletal muscle mass [8], the main source of Cr.

Since Cr is the balance of creatinine released from skeletal muscle and is removed by the kidneys, it is possible that the variables currently included in GFR prediction equations, such as age, gender, height, weight and ethnicity, are employed because they are important determinants of muscle mass [8] as well as independent predictors of kidney function. Although previous studies have used body composition to estimate kidney function [9–14], none of them have investigated the mediating relationship between the commonly used demographic/anthropometric predictor variables with Cr and GFR, which requires simultaneously investigating relationships between: (i) demographic/anthropometric variables and muscle mass; (ii) demographic/anthropometric variables and GFR; and (iii) muscle mass and GFR, using hierarchial multiple regression analyses. Thus, the first aim of this study was to test the hypothesis that muscle mass mediates a significant proportion of the variance in GFR explained by age, gender, height and weight.

If this hypothesis is true, this may help explain the inaccuracy in GFR estimation shown by previous authors. For example, Beddhu et al. [15] showed that in women of the same age and Cr, and subsequently the same GFR as predicted by the modification of diet in renal disease (MDRD) equation [2], measured 24 h creatinine excretion varied from 0.5 to 1.5 g/day. Beddhu et al. [15] further suggested that the underlying assumption of the MDRD equation (that age, gender, and ethnicity directly account for creatinine production) was inaccurate because these variables are only indirect measures of muscle mass. Thus, in chronic kidney disease patients with muscle wasting, MDRD estimated GFR is spuriously high [15].

Therefore, the second aim of this study was to investigate whether the inclusion of a direct measure of muscle mass explains additional variance in GFR than demographic and anthropometric variables alone. This is plausible because demographic and anthropometric variables alone do not take into account the individual variability in muscle mass due to genetics, diet, activity level, disease state and medication use [16], and thus may lead to inaccuracy in accounting for generation of creatinine. This study was designed to test theoretical hypotheses and not to generate a clinically applicable method of estimating GFR, which is the subject of a companion paper published elsewhere [17].



   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Acknowledgements
 References
 
Potential subjects were approached during routine pre-dialysis clinics held between September 2004 and September 2005 at Ysbyty Gwynedd and Ysbyty Glan Clwyd (prospective study, convenience sampling method). Inclusion criteria were the following: age >18 years and presence of chronic kidney disease classified as KDOQI stages 1–4. Patients with GFR >60 ml/min/1.73 m2 had to show other evidence of kidney damage: albuminuria [two spot urine samples with albumin–creatinine ratio >30 µg albumin/mg creatinine confirmed by overnight urine collections with an excretion rate of 20–200 mg albumin (microalbuminuria) or >200 mg albumin (macroalbuminuria)]: non-urological haematuria, structural abnormalities, or biopsy-proven chronic glomerulonephritis. Exclusion criteria were current dialysis or renal transplant, diabetes, conditions causing significant, uncontrolled clinically apparent oedema, use of medications affecting tubular secretion of creatinine, previous reaction to inulin, severe asthma, inability to lie flat during scanning procedures, cardiac pacemakers and inability to give written informed consent. Ethical approval was obtained from the North Central Wales Local Research Ethics Committee and all the participants gave written informed consent.

Patients presented following overnight fast on two occasions separated by a maximum of 7 days. The day before each visit, patients were requested to minimize alcohol and caffeine intake, ensure adequate hydration and avoid strenuous exercise. During the initial visit, patients attended their renal unit for kidney function assessment and blood letting. The criterion method for measuring GFR was the single shot bolus injection and total body clearance of inulin (GFRinu). Patients drank 10 ml of water per kg body weight before the procedures began and 100 ml/h thereafter. A cannula was placed in the back of each hand and baseline blood samples were drawn for subsequent analysis of Cr inulin of 2.5 g of 10 ml of 25% Inutest (Fresenius Kabi, Austria) was administered over 3 min through one cannulae before it was removed. Six samples of 4 ml were then drawn into heparinized plastic tubes from the remaining cannuale at 5, 12, 40, 120, 240 and 300 min post-injection. All samples, which were obtained after removing and discarding the initial 2 ml of blood to prevent dilution by the saline flush, were immediately centrifuged and the serum separated and stored at –80°C awaiting analysis. Inulin was assayed in plasma by a double enzyme method as previously described [18] (inter-assay CV = 4.1%; intra-assay CV = 1.2%). The reliability of single shot inulin clearances obtained in a group of 12 patients with chronic kidney disease tested on three occasions, each separated by 2 weeks, is reportedly 7.1% [19]. Cr was determined according to a modified Jaffé method using an autoanalyser (Advia 2400, Bayer, New York, USA); normal range in males = 0.9–1.3 mg/dl and in females = 0.6–1.1 mg/dl; inter-assay CV = 3.2% (n = 31); intra-assay CV = 2.9% (n = 15, at a Cr of 0.8 mg/dl). The reliability of Cr obtained in 17 CKD patients tested over 3 weeks is reportedly >18% [20].

During the second visit, patients attended the laboratories of the School of Sport, Health and Exercise Sciences at the University of Wales, Bangor for body composition assessment. Subjects voided, changed into a standard hospital gown and were barefoot. Height and weight were measured using a wall mounted stadiometer (Body Care, Warwickshire, UK) and balance scales (Seca, Hamburg, Germany), respectively. Regional and whole body lean and fat masses were assessed by dual energy X-ray absorptiometry (QDR1500, software version 5.72, Hologic, Waltham, USA). To provide a proxy measure of skeletal muscle mass, appendicular lean mass was calculated by summing the lean mass of the four limbs [21] and was expressed as a raw score or normalized by height squared (m2) to give the appendicular lean mass index. Appendicular lean mass is highly correlated with total body skeletal muscle mass [r = 0.96, standard error of the estimate (SEE) = 1.63 kg] because ~76% of the appendicular lean mass is in the muscle (the remainder being skin and connective tissue) and because ~74% of the total body skeletal muscle mass is in the extremities [21]. Fat mass was expressed as a raw score or normalized by height squared (m2, fat mass index) or as a percentage of the total weight (fat%). Since kidney size and function is related to body size, muscle and fat masses were entered in multiple regression analyses as indexes if control for scaling artifacts was required. Reliability data obtained in our lab on a group of nine patients with end-stage kidney disease tested on two occasions separated by 3 months suggests a CV for lean mass by dual energy X-ray absorptiometry of 2.4% and a standard error of measurement of 1.14 kg [22].

Statistics
All analyses were performed on a statistical computer package (SPSS version 12, Illinios, USA). Values are expressed as means ± SD. For regression, r was used to denote simple Pearson's correlation coefficients, R was used to denote accumulative coefficients from multiple regression analyses and SEE was presented when appropriate.

The principle aim of this study was to test various hypotheses regarding the relationship between Cr, GFR and muscle mass, and not to evaluate the performance of the currently used abbreviated MDRD equation, nor to generate prediction equations for use in clinical practice. We wished to avoid comparing an equation using constants generated on our data set with the MDRD equation, which uses constants generated on a different population, since performance of our equation would, of course, be better as it is specific to our subjects. Therefore, we used multiple regression analyses in an explanatory method to establish the following:

The mediating role of muscle mass between demographic/anthropometric variables and GFR
(a) Whether demographic/anthropometric variables currently included in prediction equations adequately account for variance in muscle mass (appendicular lean mass) and hence variance in creatinine production (hierarchical method); (b) whether adding a marker of muscle mass would explain extra variance in Cr than all demographic/anthropometric variables (age, gender, height and weight) (hierarchical method); and (c) whether demographic/anthropometric variables explain an independent proportion of variance in GFRinu once a marker of muscle mass had already been entered (hierarchical method).

The importance of muscle mass in explaining variance in GFR
(a) Whether adding a marker of muscle mass (appendicular lean mass) to Cr would explain extra variance in GFRinu than demographic variables currently included in the abbreviated MDRD equation alone (age and gender—ethnicity was not required as our population were all Caucasian) (hierarchical method); (b) whether adding a marker of muscle mass to Cr would explain extra variance in GFRinu than all the traditionally used demographic/anthropometric predictor variables alone (age, gender, height and weight) (hierarchical method); and (c) whether a marker of muscle mass would be selected in preference to the demographic/anthropometric variables as a predictor of GFR (stepwise method).

With multiple regression techniques, the following information can be gained: (a) the proportion of variance explained by the whole model (the R2 value, Table 2, second column, final row); (b) the proportion of variance accounted for by each predictor variable (the R2 change value, Table 2, column 3). Note: only predictors with significant F change statistics (Table 2, columns 4 and 5) significantly explain a proportion of variance in the predicted variable; and (c) the nature of the relationship, (including whether positive or negative) between the predicted and the predictor variables (standardized regression coefficients or beta values, Table 2, columns 6 and 7). In addition, the beta values detail the contribution of each individual predictor variable once all the model steps have been entered [e.g. Table 2, weight contributed to variance in GFRinu (beta = 0.220, P = 0.025) while age, gender and height did not (P = 0.622–0.975)].


View this table:
[in this window]
[in a new window]

  		Table 2. Hierarchical multiple regression analyses to investigate the mediating role of muscle mass between demographic/anthropometric variables and GFR

 
For the multiple regression analyses, assumptions of normality, linearity and homoscedasticity were visually checked by plotting the standardized residuals vs the predicted values. Transformations were made if required but clarity is detailed only when deemed necessary. Multicollinearity was checked by ensuring variance inflation factors did not exceed 100 and predictor variables did not have a condition index exceeding 100 or a condition index exceeding 30 when the predictor variable also contributed 50% of the variance on two or more regression coefficients. Influential outliers on predictor variables were identified using leverage statistics and Cook's distances, and on the predicted variable by standardized residuals >3.0 [23].



   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Acknowledgements
 References
 
Patient characteristics are detailed in Table 1. Causes of chronic kidney disease were arteriopathic (n = 11), glomerulonephritis (n = 25), infective/obstructive (n = 7), congenital/familial/hereditary (n = 5), toxicity induced (n = 2), systemic (n = 7) and uncertain (n = 20). Using predicted GFR by the MDRD equation, we attempted to recruit patients classified as KDOQI stages 1–4. However, when GFRinu was determined, one of our patients was found to be stage 5 (data not excluded). Whilst we did not specifically record blood pressure data, we aim to maintain blood pressure within guidelines as recommended by the Renal Association. The percentage of patients classified as being obese was 29.9% (as defined by a body mass index of >30 kg/m2) or 42.9% (as defined by a fat% by dual energy X-ray absorptiometry of >35% in males and >40% in females). Appendicular lean mass index correlated with weight (r = 0.678, P < 0.001), gender (r = – 0.661, P < 0.001) and height (r = 0.537, P < 0.001) but not age (r = 0.196, P < 0.087).


View this table:
[in this window]
[in a new window]

  		Table 1. Patient characteristics (n = 77)

 
Multiple regression analyses
The mediating role of muscle mass between demographic/anthropometric variables and GFR
  1. To determine whether demographic variables (age, gender, height and weight) adequately account for variance in muscle mass (appendicular lean mass), these variables were forced into a hierarchical regression model (Table 2). Although a significant model was generated (R2 = 0.828, F(4,72) = 86.46, P < 0.001) and beta values revealed all entered variables explained a significant proportion of variance in appendicular lean mass, 17.2% of the variance in appendicular lean mass was still unaccounted for. Furthermore, the SEE of the estimate of appendicular lean mass prediction was relatively large (2.33 kg).
  2. To determine the effect including a measure of body composition has on explaining variance in Cr, the following variables were forced into a hierarchical regression model: GFRinu, age, gender, height and weight plus muscle mass (appendicular lean mass) (R2 = 0.648, F(6,70) = 21.47, P < 0.001) (Table 2). Forcing in appendicular lean mass explained an additional 4.2% of the variance in Cr. Analysis of beta values revealed only appendicular lean mass and GFRinu significantly contributed to variance in Cr.
  3. To determine whether the traditionally used demographic variables would explain extra variance in GFRinu to that explained by muscle mass (appendicular lean mass index), these variables were entered after Cr and appendicular lean mass index into a hierarchical model (Table 2). Only weight explained additional variance (2.0%) in GFRinu.

The importance of muscle mass in explaining variance in GFR

  1. To determine the value of adding a measure of body composition to the current MDRD method to predict GFRinu, a hierarchical regression model was generated that included the variables used in the abbreviated MDRD equation (Cr, age and gender) plus muscle mass (appendicular lean mass index) (R2 = 0.700, F(4,72) = 42.09, P < 0.001, Table 3). Forcing in appendicular lean mass index after the MDRD variables explained a further 10.4% of the variance in GFRinu.
  2. To determine the value of including a measure of body composition to all the demographic/anthropometric variables regularly used to predict GFR, a hierarchical regression model was generated that included Cr and all the traditionally used demographic/anthropometric variables (age, gender, height and weight) plus muscle mass (appendicular lean mass index) (R2 = 0.731, F(6,70) = 31.75, P < 0.001, Table 3). Forcing in appendicular lean mass index after all the other potential variables explained an additional 2.4% of the variance in GFRinu. Beta values revealed only Cr, appendicular lean mass index and weight significantly contributed to variance in GFRinu. Forcing in fat mass index or fat% did not account for extra variance in GFRinu (data not shown).
  3. To determine which variables were most strongly associated with GFRinu, a stepwise regression model was generated that selected Cr, muscle mass (appendicular lean mass) and weight (R2 = 0.729, F(3,73) = 65.426, P < 0.001, Table 3). Excluded variables were height (P = 0.310), age (P = 0.725), gender (P = 0.772) and fat mass (P = 0.845). The following predictive equation was generated:


Formula

where GFRinu, log 10 transformed glomerular filtration rate by inulin clearance (ml/min); log 10transformed serum creatinine concentration (mg/dl); ALM, appendicular lean mass; weight, square root transformed weight (kg). The correlation between GFRinu and predicted GFR using appendicular lean mass was r = 0.882 (P < 0.001); the SEE of the estimate was 16.7 ml/min.


View this table:
[in this window]
[in a new window]

  		Table 3. Multiple regression analysis explaining variance in glomerular filtration rate

 


   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Acknowledgements
 References
 
The present study aimed to test whether muscle mass mediates the relationship between demographic/anthropometric variables and GFRinu. Using serum creatinine-based measures of kidney function, the relationships between these variables have been extensively researched [5–7]. However, while many studies admit results may be confounded by altered muscle mass, to the best of our knowledge, no studies have simultaneously taken into account a measure of body composition and used hierarchical multiple regression analyses to investigate the mediating effect of muscle mass on the relationship between demographic/anthropometric variables and GFR.

Firstly, as has been previously speculated [15], we empirically showed that demographic/anthropometric variables do not adequately account for variance in muscle mass. In the present data set, demographic variables (age, gender, height and weight) left 17.2% of the variance in appendicular lean mass unexplained, resulting in a relatively large SEE of the appendicular lean mass prediction (2.33 kg). Performance of the demographic variables as used in the abbreviated MDRD equation (just age and gender) was even worse. This is because age, gender, height and weight alone fail to account for other factors that can influence body composition (namely genetics, activity level, diet, disease state and medication use [16]). Thus relying on demographic/anthropometric variables alone rather than measuring muscle mass directly could lead to errors when accounting for creatinine production and consequently error in GFR estimation.

In accordance with this hypothesis, our data revealed that demographic/anthropometric variables also failed to completely account for variations in Cr. More than 90% of the total body creatine is found in skeletal muscle, wherein an almost constant conversion of creatine to creatinine occurs (approximately 2 g a day assuming a total of 120 g of creatine in a 70 kg man) which then diffuses from muscle to blood. Theoretically, directly measuring muscle mass will account for more variance in creatinine production. The present data set supports this hypothesis: by adding a marker of muscle mass to GFRinu after accounting for all the commonly used demographic/anthropometric variables (i.e. age, gender, height and weight), an extra 4.2% of the variance in Cr was explained.

Furthermore, we built a hierarchical regression model to predict GFRinu with Cr and muscle mass being entered first. Of the demographic/anthropometric variables subsequently forced in, only weight significantly added to the variance explained in GFRinu. Age did not explain additional variance in GFRinu. While it is generally accepted that GFR declines with age, our data suggest that the relationship of Cr-estimated GFRinu and age is reduced once the variations in muscle mass are accounted for, and that the relationship between Cr-estimated GFR and age is primarily due to the reduction in muscle mass seen with age [16]. In fact, despite a reduction in kidney function, Cr may remain normal due to reduced creatinine generation [6]. Thus, relying on a creatinine based estimation of GFR may be inaccurate if muscle mass is not taken into account.

A similar finding was observed for gender. While female gender is generally believed to be protective against GFR decline (possibly due to hormonal differences) [7], again our data suggest that the relationship of Cr-estimated GFRinu with gender is reduced once the variations in muscle mass are accounted for. As some females may have higher muscle mass than some males, relying on a creatinine based estimation of GFR and applying a crude correction factor to account for differences in gender will be inaccurate if muscle mass is not directly measured.

However, weight still accounted for some variance in GFRinu once muscle mass had been entered. Weight most likely was entered because our appendicular lean mass derived marker of total body skeletal muscle did not take into account the muscle in the trunk region (which weight would include) or because of a scaling artifact between body size (which is correlated with weight) and kidney size/function. Another potential explanation is the hyperfiltration observed in overweight persons [5].

Taken together, these findings suggest that muscle does indeed mediate the relationship between demographic/anthropometric variables and Cr-estimated GFR. Thus, our findings provide empirical evidence explaining why demographic/anthropometric variables have been entered in previous equations such as the MDRD formulae, i.e. variables such as age and gender may not only be directly related to kidney function, but primarily correlate with Cr-estimated GFR because muscle mass declines with age and varies with gender. As the prediction of muscle mass using demographic and anthropometric variables is poor, we hypothesized that more variance in GFR would be accounted for if muscle mass was measured directly rather than predicting it.

The second aim of the present study was to test this hypothesis. We found that by adding muscle mass to the demographic variables used in the currently advocated abbreviated MDRD equation (i.e. age and gender), a significant and substantial improvement in the amount of the variance accounted for in measured GFRinu was obtained (10.4%). This is a conservative estimate of improvement because we reduced scaling artifact by entering muscle mass as an index of height; and because we generated our own constants for use with the MDRD equation demographic variables (rather than using actual MDRD generated constants which perform poorly in independent samples such as ours). Even adding a marker of muscle mass to all the commonly used demographic/anthropometric predictor variables (i.e. age, gender, height and weight) still significantly increased the explained variance in GFRinu by 2.4%. Furthermore, using stepwise regression, muscle mass was selected in preference to demographic/anthropometric variables to explain variance in GFRinu.

Thus, these findings provide support for the hypothesis that including a measure of muscle mass will explain more variance in GFR than anthropometric/demographic variables alone. However, it should also be noted that including weight also explained additional variance in GFRinu, albeit to a lesser degree than muscle mass, and albeit possibly causing GFR estimation inaccuracy in patients with abnormal body composition [because of no differentiation between muscle (creatinine generating) and fat (non-creatinine generating) mass]. Finding that weight explained additional variance in GFR is important because the MDRD equation (which does not include anthropometric data) has largely replaced the Cockcroft and Gault equation (that includes weight).

Limitations of the present study include reliance on previously published data for determining reliability and validity of our inulin and creatinine methods. Specifically, total body clearance of inulin (as utilized in the present study) vs urinary clearance may yield higher values. In addition we did not prescribe a meat-free diet preceding the study, possibly reducing reliability of Cr. There may also be error associated with using dual energy X-ray absorptiometry to measure muscle mass when hydration status is altered.

In the present study, we did not wish to develop and validate a prediction equation for estimating GFR. Rather, we wished to test the hypothesis that muscle mass mediates the relationship between demographic/anthropometric variables and GFR; a hypothesis that our data support. Nevertheless, using a stepwise regression model, an equation was generated utilizing Cr, appendicular lean mass by dual energy X-ray absorptiometry and weight that could be used to estimate GFR. A logical avenue for future research is to develop clinically obtainable measures to predict appendicular lean mass and to determine whether using predicted appendicular lean mass can improve GFR prediction compared with currently advocated GFR equations. Answering this question forms the basis of our companion paper [17].

In summary, we provide evidence to support the hypothesis that muscle mass mediates the relationship between demographic/anthropometric variables and creatinine-estimated GFR. Consequently, including a measure of muscle mass explained additional variance in GFR than using anthropometric/demographic variables alone. It is probable that by measuring muscle mass, accuracy of GFR estimation equations could be improved, especially in patients with abnormal body composition.



   Acknowledgements
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Acknowledgements
References
 
This study was supported by a start up grant from Kidney Research UK (RP34/1/2004). Our research group receives a proportion of its funding from local NHS executives.

Conflict of interest statement. None declared.

(See related article by Macdonald et al. Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in non-diabetic patients with chronic kidney disease. Nephrol Dial Transplant 2006; 21: 3481–3487.)



   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Acknowledgements
 References
 

  1. Tett SE, Kirkpatrick CMJ, Gross AS, McLachlan AJ. (2003) Principles and clinical application of assessing alterations in renal elimination pathways. Clin Pharmacokinet 42:1193–1211.[CrossRef][ISI][Medline]
  2. Levey AS, Bosch JP, Lewis JB, et al. (1999) A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 130:461–470.[Abstract/Free Full Text]
  3. Cockcroft DW and Gault MH. (1976) Prediction of creatinine clearance from serum creatinine. Nephron 16:31–41.[ISI][Medline]
  4. Levey AS, Greene T, Kusek J, Beck GJ, Group MS. (2000) A simplified equation to predict glomerular filtration rate from serum creatinine (Abstract). J Am Soc Nephrol 11:A0828.
  5. Ribstein J, du Cailar G, Mimran A. (1995) Combined renal effects of overweight and hypertension. Hypertension 26:610–615.[Abstract/Free Full Text]
  6. Macias-Nunez JF and Cameron JS. (2005) The ageing kidney. In Davison AM, Cameron JS, Grunfeld J (Eds.), et al. Oxford Textbook of Clinical Nephrology.(Oxford University Press, Oxford) pp. 73–86.
  7. Eriksen BO and Ingebretsen OC. (2006) The progression of chronic kidney disease: A 10-year population-based study of the effects of gender and age. Kidney Int 69:375–382.[CrossRef][ISI][Medline]
  8. Lee RC, Wang Z, Heo M, et al. (2000) Total-body skeletal muscle mass: development and cross validation of anthropometric prediction models. Am J Clin Nutr 72:796–803.[Abstract/Free Full Text]
  9. Taylor TP, Wang W, Shrayyef MZ, et al. (2006) Glomerular filtration rate can be accurately predicted using lean mass measured by dual-energy X-ray absorptiometry. Nephrol Dial Transplant 21:84–87.[Abstract/Free Full Text]
  10. Donadio C, Lucchesi A, Tramonti G, Bianchi C. (1997) Creatinine clearance predicted from body cell mass is a good indicator of renal function. Kidney Int 63:S166–S168.
  11. Donadio C, Lucchesi A, Tramonti G, Bianchi C. (1998) Creatinine clearance can be predicted from plasma creatinine and body composition analysis by means of electrical bioimpedance. Ren Fail 20:285–293.[ISI][Medline]
  12. Donadio C, Consani C, Ardini M, Caprio F, Grassi G, Lucchesi A. (2004) Prediction of glomerular filtration rate from body cell mass and plasma creatinine. Curr Drug Discov Tech 1:221–228.
  13. Sanaka M, Takano K, Shimakura K, Koike Y, Mineshita S. (1995) Rapid and accurate estimation of creatinine clearance in the muscle-wasted elderly by computed tomography. Gerontology 41:332–342.[ISI][Medline]
  14. Lim WH, Lim EM, McDonald S. (2006) Lean body mass-adjusted Cockcroft and Gault formula improves the estimation of glomerular filtration rate in subjects with normal-range serum creatinine. Nephrology 11:250–256.[CrossRef][Medline]
  15. Beddhu S, Samore MH, Roberts MS, et al. (2003) Creatinine production, nutrition, and glomerular filtration rate estimation. J Am Soc Nephrol 14:1000–1005.[Abstract/Free Full Text]
  16. Mitchell D, Haan MN, Steinberg FM, Visser M. (2003) Body composition in the elderly: the influence of nutritional factors and physical activity. J Nutr Health Aging 7:130–139.[Medline]
  17. Macdonald JH, Marcora SM, Jibani MM, et al. (2006) Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in patients with chronic kidney disease. Nephrol Dial Transplant 21:3481–3487.[Abstract/Free Full Text]
  18. Soper CP, Bending MR, Barron JL. (1998) Long-term glycaemic control directly correlates with glomerular filtration rate in early Type 1 diabetes mellitus before the onset of microalbuminuria. Diabet Med 15:1010–1014.[CrossRef][ISI][Medline]
  19. Florijn KW, Barendregt JNM, Lentjes EGWM, et al. (1994) Glomerular filtration rate measurement by single-shot injection of inulin. Kidney Int 46:252–259.[ISI][Medline]
  20. Holzel WG. (1987) Intra-individual variation of some analytes in serum of patients with chronic renal failure. Clin Chem 33:670–673.[Abstract/Free Full Text]
  21. Kim J, Wang Z, Heymsfield SB, Baumgartner RN, Gallagher D. (2002) Total-body skeletal muscle mass: estimation by a new dual-energy X-ray absorptiometry method. Am J Clin Nutr 76:378–383.[Abstract/Free Full Text]
  22. Macdonald JH, Marcora SM, Jibani M, et al. (2005) Intradialytic exercise as anabolic therapy in haemodialysis patients – a pilot study. Clin Physiol Funct Imaging 25:113–118.[CrossRef][ISI][Medline]
  23. Stevens JP. (2002) Applied Multivariate Statistics for the Social Sciences.(Lawrence Erlbaum Associates, London).
  24. Du Bois D and Du Bois EF. (1916) A formula to estimate the approximate surface area if height and weight are known. Arch Intern Med 17:863–871.[ISI]
Received for publication: 13. 2.06
Accepted in revised form: 22. 6.06


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?

Related articles in NDT:

Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in non-diabetic patients with chronic kidney disease
Jamie H. Macdonald, Samuele M. Marcora, Mahdi Jibani, Gareth Roberts, Mick John Kumwenda, Ruth Glover, Jeffrey Barron, and Andrew Bruce Lemmey
NDT 2006 21: 3481-3487. [Abstract] [FREE Full Text]  



This article has been cited by other articles:


Home page
 Nephrol Dial TransplantHome page
J. H. Macdonald, S. M. Marcora, M. Jibani, G. Roberts, M. J. Kumwenda, R. Glover, J. Barron, and A. B. Lemmey
Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in non-diabetic patients with chronic kidney disease
Nephrol. Dial. Transplant., December 1, 2006; 21(12): 3481 - 3487.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
21/12/3488    most recent
gfl430v2
gfl430v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Related articles in NDT
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Macdonald, J. H.
Right arrow Articles by Lemmey, A. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Macdonald, J. H.
Right arrow Articles by Lemmey, A. B.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?