 Research article
 Open Access
 Open Peer Review
 Published:
Differences in estimation of creatinine generation between renal function estimating equations in an Indian population: crosssectional data from the Hyderabad arm of the Indian migration study
BMC Nephrology volume 14, Article number: 30 (2013)
Abstract
Background
Creatinine based formulae for estimating renal function developed in white populations may be less valid in other ethnic groups. We assessed the performance of various estimating formulae in an Indian population.
Methods
917 subjects were recruited from the Hyderabad arm of the Indian Migration Study. Data were collected on comorbidity, serum creatinine and body composition from DXA scans. Renal function was compared using the modified CockcroftGault, MDRD and CKDEPI formulae. 24hour creatinine production was derived from each estimate and the agreement with measured muscle mass examined. 24hour creatinine production estimates were compared to that derived from a formula by Rule incorporating DXA measured muscle mass. Potential systematic biases were examined by age and eGFR. We assessed the association of renal function by each formula with hypertension and selfreported measures of vascular disease.
Results
Mean modified CockcroftGault eCCl was 98.8 ml/min/1.73 m^{2}, MDRD eGFR 91.2 ml/min/1.73 m^{2} and CKDEPI eGFR 96.3 ml/min/1.73 m^{2}. MDRD derived 24hour creatinine production showed the least agerelated underestimation compared to the Rule formula. CKDEPI showed a marked bias at higher eGFRs. All formulae showed similar strength associations with vascular disease and hypertension.
Conclusions
Our analyses support the use of MDRD for estimating renal function in Indian populations. Further work is required to assess the predictive value of formulae for incident disease and complications of CKD.
Background
Chronic kidney disease (CKD) is increasingly recognised as a public health problem, associated with an increased risk of cardiovascular disease, EndStage Renal Failure (ESRF) requiring renal replacement therapy (RRT) and a reduced life expectancy [1]. In people with CKD, cardiovascular disease is the leading cause of death, and CKD is an independent risk factor for coronary heart disease (CHD) [2]. In the UK, crude relative acceptance rates for RRT in South Asians (from India, Pakistan and Bangladesh) is 3.5 times that seen in Whites [3]. The burden of CKD in India is not quantified but given the burden of diabetes, hypertension and cardiovascular disease, it is predicted to be high [4]. Information on CKD is needed to enable intervention at early stages of disease to prevent complications and progression to ESRF.
All creatininebased formulae used to estimate renal function imply an estimation of creatinine generation rate based on demographic variables. It is recognised that methods of estimating Creatinine Clearance (CCl) and Glomerular Filtration Rate (GFR) developed in White populations may require adjustment for use in other ethnic groups [5, 6]. AfricanAmerican individuals have a higher generation of creatinine from muscle than White subjects, due mostly to greater skeletal muscle mass [7, 8]. Thus coefficients for AfricanAmericans within MDRD and CKDEPI result in higher values for estimated GFR (eGFR) than Whites for the same plasma creatinine concentration, age and gender [9]. Similar differences in skeletal muscle mass and creatinine generation exist between Indian and White populations [10]. To date, none of the renal function estimating formulae has been well validated in an Indian population. The CKDEPI research group carried out work on developing an ‘Asian’ coefficient but this aimed to incorporate primarily Chinese and Japanese populations [5].
Rule and colleagues recently investigated the use of Dualenergy Xray absorptiometry (DXA) measurements of muscle mass to estimate Urinary Creatinine Clearance (UCCl) and compared this to the use of demographic variables [11]. A model based on muscle mass and age had the best fit.
In this mixed urbanrural Indian population from Hyderabad we investigated different methods of estimating renal function, including Rule’s formula using measured muscle mass. The objective of this study was to validate the estimation of creatinine generation between different equations used to estimate GFR or CCl against DXA measured muscle mass in an Indian population. We derived 24hour (24 h) creatinine production from each renal function estimate and compared this to measured muscle mass. This comparison provides a unique insight into the ability of demographic variables within each formula to estimate creatinine generation in an Indian population. Demographic variables in creatininebased estimating formulae primarily account for differences in muscle mass, from which creatinine is largely derived.
We compared estimates of excretory renal function, and consequent classification of CKD, by different formulae. Further, we assessed the association of selfreported measures of vascular disease and hypertension with each estimate.
Methods
Study methodology
Using the framework of a cardiovascular risk factor screen study conducted in factories in north, central and south India [12] a sibpair comparison study was designed. Details of this are reported elsewhere [13]. The original study was in four factories: Lucknow, Hindustan Aeronautics Ltd; Nagpur, Indorama Synthetics Ltd; Hyderabad, Bharat Heavy Electricals Ltd; and Bangalore, Hindustan Machine Tools Ltd. Factory workers and their coresident spouses were recruited if they were ruralurban migrants using employer records as the sampling frame. Each migrant worker and spouse was asked to invite one nonmigrant samesex fullsibling, closest to them in age, residing in their rural place of origin. A 25% random sample of nonmigrants was invited to participate in the study. Nonmigrants were also asked to invite a sibling who resided in the same city but didn’t work in the factory. Information sheets were translated into local languages and signed (thumb print acceptable) and informed consent obtained. Ethics approval was from the All India Institute of Medical Sciences Ethics Committee, reference no. A60/4/8/2004. Field work ran from March 2005 to December 2007. The Hyderabad arm of the Indian Migration Study is a followup study of the Hyderabad recruited subjects. The participants were invited to attend a screening clinic at the National Institute of Nutrition between January 2009 and December 2010.
Clinic measures
Participants were interviewed and data collected on demographic factors, including socioeconomic status using a subset of 12 questions from the Standard of Living Index (SLI), a household level assetbased scale devised for Indian surveys [14]. These comprised house type, house ownership, toilet facility, lighting source, drinking water source, car/tractor, scooter, telephone, refrigerator, television, bicycle, radio, clock/watch, and weighted to give a maximum score of 33. Weighting was developed by the International Institute of Population Sciences. Individuals were classed as having ‘low’ (07), ‘middle’ (812) or ‘high’ (1333) standards of living. Smoking was assessed as positive if individuals actively smoked or chewed tobacco. Few individuals reported past but not active smoking. A diagnosis of CHD and Stroke was made by selfreport of a doctor diagnosis. Weight was measured twice without shoes using digital Seca scales (http://www.seca.com). Standing height was measured twice without shoes using a portable stadiometer (Leicester height measure; Chasmors Ltd, Camden, London, UK). Body Mass Index was calculated as weight(kg)/height(m)^{2}. We used a validated oscillometric device (OMRON M5I; Omron, Matsusaka Co, Japan) to measure blood pressure (BP) in the sitting position with appropriate cuff sizes. We took three measures 23 minutes apart, and averaged the last two measures for analyses. A diagnosis of hypertension was made if average systolic BP was ≥140 mmHg, average diastolic BP ≥90 mmHg or if there was report of a doctor diagnosis of hypertension [15].
Participants were asked to attend fasting and the time of their last meal was recorded. Creatinine analysis was performed using the rateblanked compensated Jaffe method on a Roche COBASC311 autoanalyzer. The calibraton for this assay is traceable to isotope dilution mass spectrometry (IDMS). The Cardiac Biochemistry Lab, AIIMS, is part of the UK National External Quality Assessment Scheme (http://www.ukneqas.org.uk) to quality assure assays.
Whole body DXA scans were performed on a HologicDXA machine DiscoveryA model (91% scans) or a HologicQDR4500Elite machine (http://www.hologic.com) (9% scans). Scans were individually checked for major artefacts and if present were removed from the analyses. As quality assurance, a spine phantom was scanned daily to check for acceptable ranges. Total skeletal muscle mass was determined from the lean body mass of all four extremities multiplied by 1.33 [16]. Body Surface Area was calculated by the Mosteller equation [17]. Renal function was estimated using four measures, including Rule’s formula incorporating measured muscle mass [11]:
Renal function estimating equations:

1.
Creatinine Clearance by CockcroftGault modified for Body Surface Area (BSA) by the Mosteller equation.
$$\begin{array}{l}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{eCCl}}=\left(140\u2013\mathrm{\text{Age}}\right)\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{Mass}}\left(\mathrm{\text{kg}}\right)\\ \phantom{\rule{4em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\left[0.85\phantom{\rule{0.25em}{0ex}}\mathrm{if}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{female}}\right]/\text{serum creatinine (mg/dl)}\\ \phantom{\rule{2.7em}{0ex}}=\mathrm{\text{ml}}/\mathrm{\text{min}}\end{array}$$Serum creatinine (mg/dl)
$$\begin{array}{l}\mathrm{\text{Modified}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{for}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{BSA}}=\mathrm{\text{eCCl}}\phantom{\rule{0.25em}{0ex}}\mathrm{\times}1.73/\text{BSA}\\ \phantom{\rule{8.7em}{0ex}}=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}$$BSA

2.
The IDMS standardized Modification of Diet in Renal Disease (MDRD) formula
$$\begin{array}{l}\mathrm{\text{eGFR}}=175\phantom{\rule{0.25em}{0ex}}\mathrm{\times}(\phantom{\rule{0.25em}{0ex}}\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}\phantom{\rule{0.25em}{0ex}}{\left(\mathrm{\text{mg}}/\mathrm{\text{dl}}\right)}^{1.154}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{Ag}}{\mathrm{e}}^{0.203}\\ \phantom{\rule{4.5em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\left[0.742\phantom{\rule{0.25em}{0ex}}\mathrm{\text{if}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{female}}\right]=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}$$ 
3.
The CKDEPI formulae
For ‘nonblack females’:
If Serum creatinine ≤ 0.7 mg/dl
$$\begin{array}{l}\mathrm{\text{eGFR}}=144\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}{\left(\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}/0.7\right)}^{0.329}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}0.99{3}^{\mathrm{\text{Age}}}\\ \phantom{\rule{2.7em}{0ex}}=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}$$If Serum creatinine > 0.7 mg/dl
$$\begin{array}{l}\mathrm{\text{eGFR}}=144\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}{\left(\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}/0.7\right)}^{1.209}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}0.99{3}^{\mathrm{\text{Age}}}\\ \phantom{\rule{2.7em}{0ex}}=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}$$For ‘nonblack males’:
If Serum creatinine ≤0.9 mg/dl
$$\begin{array}{l}\mathrm{\text{eGFR}}=141\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}{\left(\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}/0.9\right)}^{0.411}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}0.99{3}^{\mathrm{\text{Age}}}\\ \phantom{\rule{2.7em}{0ex}}=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}$$If Serum creatinine >0.9 mg/dl
$$\begin{array}{l}\mathrm{\text{eGFR}}=141\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}{\left(\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}/0.9\right)}^{1.209}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}0.99{3}^{\mathrm{\text{Age}}}\\ \phantom{\rule{2.7em}{0ex}}=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}$$ 
4.
Estimated creatinine clearance based on Rule formula DXA measured skeletal muscle mass and age
If age > 55 years:
$$\frac{\begin{array}{ccc}\mathrm{\text{eCrClagemuscle}}& =& \mathrm{\text{exp}}\phantom{\rule{0.5em}{0ex}}\left[6.37+0.029\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\left(\mathrm{\text{muscle}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{in}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{kg}}/1.73{\mathrm{m}}^{2}\right)\right.\\ \left.\left[0.0071\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{age}}55\right]\right]=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\end{array}}{\left(\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}\phantom{\rule{0.25em}{0ex}}\left(\mathrm{\text{mg}}/\mathrm{\text{dl}}\right)\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}14.4\right)}$$If age ≤55 years:
$$\frac{\begin{array}{ccc}\mathrm{\text{eCrClagemuscle}}\hfill & =\hfill & \mathrm{exp}\phantom{\rule{0.25em}{0ex}}[6.37+0.029]\hfill \\ \phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}\left(\mathrm{\text{muscle}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{in}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{kg}}/1.73{\mathrm{m}}^{2}\right)=\mathrm{\text{ml}}/\mathrm{\text{min}}/1.73{\mathrm{m}}^{2}\hfill \end{array}}{\left(\mathrm{\text{Serum}}\phantom{\rule{0.25em}{0ex}}\mathrm{\text{creatinine}}\phantom{\rule{0.25em}{0ex}}\left(\mathrm{\text{mg}}/\mathrm{\text{dl}}\right)\phantom{\rule{0.25em}{0ex}}\mathrm{\times}\phantom{\rule{0.25em}{0ex}}14.4\right)}$$
CKD was classified according to the KDOQI guidelines [18]. As information on proteinuria was unavailable, CKD 1 and 2 were not identified. We defined an eGFR <60 ml/min/1.73 m^{2} (CKD stages 35) as ‘CKD’ and an eGFR ≥60 ml/min/1.73 m^{2} as ‘No CKD’. In clinical practice, the diagnosis of CKD requires evidence of reduced function for ≥3 months; in epidemiological studies a single estimate of GFR is accepted [19].
All creatininebased GFR formulae use the serum creatinine concentration and demographic variables to empirically derive an estimate of GFR. Serum creatinine concentration is determined by the arithmetical relationship between creatinine generation rate and glomerular filtration rate, whilst noting that CCl is not precisely the same as GFR, particularly when GFR is low, because of tubular secretion. Therefore the formulae are, in effect, different ways of estimating creatinine generation rate from demographic characteristics, estimating GFR from this together with the serum creatinine concentration. Creatinine generation rate should then be:
Creatinine generation rate (mmol/24 h) = CCl or GFR (ml/min) × serum creatinine (mmol/ml) × 1440 (min)
Given that creatinine generation rate is proportional to muscle mass, one would expect to find a linear correlation between the two. The tightness of this correlation would be a measure of the accuracy of each formula.
Statistical analyses
Linear regression and Pearson’s correlation coefficient were used to investigate the relationship between each estimate of renal function and DXA measured skeletal muscle mass. BlandAltman plots, which plot the difference (delta) between two formulae against the average, were used to visually compare the relative performance of each formula. To assess how each formula may perform across a range of ages, we regressed the difference in 24 h creatinine production as estimated by the DXA based Rule formula and the traditional formulae, against age to examine for any systematic bias (increasing/decreasing differences with age). We tested for any differences in the regression coefficients using a heterogeneity test.
All creatininebased renal function formulae underperform systematically at high GFRs. Renal function estimates therefore underwent log_{2}transformation to allow this error to be more normally distributed. Therefore, one unit increase change is equivalent to doubling renal function. We used logistic regression to assess the association between the transformed renal function estimates and our disease endpoints, vascular disease and hypertension, both unadjusted and adjusted for age. We undertook sexstratified and combined analyses after formally testing for any sex interaction.
The study included related individuals who are not truly independent, so we used robust standard errors in our regression models to allow for any family clustering effect. This provides wider 95% confidence intervals than conventional methods. Statistical analyses were performed using Stata 11.
Results
The study sample comprised 917 adult subjects (mean age 48.5 ± SD 8.3 years) and roughly equal numbers of men and women (Table 1). The crude mean muscle mass in women was 7.8 kg (95% CI 7.77.9) less than that in men. However, the BMI of women was 2.2 kg/m^{2} higher than that of men (95% CI 2.12.3). The mean eCCl by mCG was 98.8 ml/min/1.73 m^{2} and by Rule’s formula 103.2 ml/min/1.73 m^{2}. eGFR as estimated by MDRD was 91.2 ml/min/1.73 m^{2} and by CKDEPI 96.3 ml/min/1.73 m^{2}. Generally, estimates of excretory renal function were higher in women than in men. In men the Rule formula generated the highest estimate of renal function, followed by CKDEPI, mCG and MDRD. In women, the findings were slightly different with mCG generating a higher estimate than CKDEPI.
In the total population the estimated prevalence of CKD stages 35 showed up to 10 fold variability; the highest was with mCG (6.7%) followed by MDRD (4.4%), CKDEPI (2.0%) and finally Rule formula (0.6%). In general, prevalence was higher in men compared to women (Table 1 and Additional file 1: Figure S1).
We examined whether there was any systematic bias between the different formulae according to the level of renal function. The BlandAltman plots (Additional file 1: Figure S3) showed that the variation in the differences between formulae gets larger as the mean level of renal function increases but in a nonspecific fashion except for CKDEPI. The CKDEPI formula shows a biased pattern: underestimating renal function compared to the other formulae at the high end of function.
Association between estimated creatinine production and muscle mass
Overall, lean muscle mass explained a greater proportion of the variance in 24 h creatinine production as estimated by MDRD, though this was not the case in the gender specific analyses. Overall and in the genderspecific analyses, the strength of the relationship, as measured by the regression gradient, was strongest with mCG. The strongest correlation was seen with the Rule formula, which was unsurprising as this was itself derived against DXA measures (Additional file 1: Table S1 and Additional file 1: Figure S3).
Formula estimated adjustment of muscle mass with age
As the Rule formula was most closely related to measured muscle mass it was used as the closest estimate of 24 h creatinine production. The difference between 24 h creatinine production as estimated by the other formulae and that by Rule was regressed against the subjects’ age. All formulae underestimated the 24 h creatinine production when compared to the Rule formula. Both mCG and CKDEPI showed a marked negative association with age (regression coefficients per year increase –0.072 (95% CI –0.080 to –0.065) and –0.035 (95% CI –0.044 to –0.027) respectively so that they underestimate 24 h Cr production more in the old than the young (Figure 1). This was less marked for the MDRD formula (regression coefficients per year increase –07.012 (95% CI –0.021 to –0.003) (formal heterogeneity test between coefficients p<0.001).
Association with vascular disease and hypertension
All formulae showed an inverse association with combined cardiac and cerebrovascular disease so that disease status was associated with worse renal function (Table 2). This was, in general, stronger for men than women. After adjustment for age these associations were attenuated; CKDEPI had the strongest inverse association but the wide 95% confidence intervals show there is no strong evidence that one formula is superior to the others. For hypertension there was an inverse association in the crude analysis with a stronger association for women. The CKDEPI formula had the strongest association but again overlapping 95% confidence intervals. After age adjustment the association attenuated so that they were consistent with chance. Differences in the effect estimates by gender were consistent with chance (all pvalues for interaction >0.05).
Discussion
This is a large study of renal function in a mixed urbanrural Indian population. The comparison with measured muscle mass provides an insight into the ability of demographic variables to estimate creatinine generation in Indian populations. This is important as the demographic variables used as surrogates for variability in creatinine production and excretion were developed in White populations and thus may not be generalisable to other ethnicities.
The mean eCCl derived from the Rule formula in healthy White Americans was 104±26 ml/min/1.73 m^{2} as compared to 103.2±20.4 ml/min/1.73 m^{2} in our Indian population [11]. It is interesting to note that the Indian population in this study had a lower mean total body mass and a lower BMI compared to the Rule’s population, but that muscle mass was comparable (Additional file 1: Table S2) showing that this Indian group is relatively lean. However, our population contained both rural and urban residents, and with a gradient of increasing obesity with urbanization recognized [20] these results should not be generalised to urban Indian populations.
Of the traditional GFR estimating equations, MDRD derived 24 h creatinine production was most strongly correlated with DXA measured muscle mass. MDRD was the only formula which did not have a marked differential bias with age and thus may be the most suitable at predicting agerelated differences in muscle mass and creatinine production. The mCG formula generates the highest mean clearance but also gives the highest proportion of those with CKD 35, which at first sight appears paradoxical. This is explained by the distribution of creatinine clearance which appears to have a larger right sided tail using mCG (Additional file 1: Figure S1).
The Rule formula led to the highest estimate of GFR, suggesting that the other formulae underestimate true muscle mass in this population. Importantly, the estimate of the prevalence of CKD 35 varies from 0.6% to 6.7% depending on which formula is used. This highlights how comparisons of studies estimating the prevalence of CKD are limited if different formulae are used and has important implications for health service planning. We didn’t find evidence that one formula was better than the others in predicting vascular disease and hypertension.
This study is the largest comparison of different estimates of creatinine clearance in an Indian population which also has DXA derived measures of muscle mass. It does, however, have several important limitations. (1) Lack of a gold standard. We did not compare the calculated GFR and CCl estimates with a ‘measured gold standard’. However, the assumption that measured GFR is the ‘gold standard’ against which all estimating formulae should be compared has been called into question [21]. There’s no universally recognised gold standard method of measuring GFR and different methods often yield different measurements. Measured GFR has not been shown to be better than other markers of renal disease at predicting outcomes associated with poor renal function [22]. Determining with great accuracy ‘actual’ GFR is less a priority than calculating a person’s risk of clinical complications. (2) CCl overestimates GFR GFR estimating equations have been developed to estimate GFR against measured GFR and underestimate measured CCl by approximately 15% [23]. CG was developed to estimate CCl by regressing UCCl/kg body weight onto age [24]. Direct comparison of CCl and GFR is therefore limited but in clinical practice they are used similarly. (3) Estimating Muscle Mass from DXA scans Wang et al’s original work, concluding that appendicular muscle mass constitutes approximately 75% of total body skeletal muscle mass [16], was on a sample of 25 White men. In further work ethnicity was not found to be a significant predictor of skeletal muscle, but of the 414 study subjects, only 0.12% were ‘Asian’ [25]. We applied Wang’s estimate to an Indian group, although lacking validation data in this population. However, any error should apply equally to all the compared formulae. (4) Contribution of diet and metabolism Creatinine is mainly produced by muscle breakdown, largely determined by muscle mass but affected by dietary protein and metabolic rate. The effect of dietary protein is minimised by ensuring serum creatinine is measured on a fasting blood sample [26], as in this study. Metabolism was not examined. (5) Comorbidity measures There may have been underascertainment of vascular endpoints in rural areas, compared to urban which benefit from employment health facilities. This may have attenuated true associations.
This study supports previous evidence that formulae to estimate renal function based on serum creatinine and demographic characteristics are all imprecise due to the impossibility of predicting steadystate creatinine generation in an individual. The MDRD formula appears less variable and less biased by age in the prediction of creatinine generation. However there was no convincing evidence that any one measure was superior in predicting complications of CKD; this may be due to a low prevalence of complications in this relatively young, healthy population. Followup will show which estimate of renal function is most strongly associated with comorbidities. Further work is required to determine the most useful estimate of renal function in an Indian population. Given the large and increasing burden of diabetes and obesity in India, this will be of major future importance in planning future renal care services.
Conclusions
Our analyses support the use of MDRD for estimating renal function in Indian populations. Further work is required to assess the predictive value of formulae for incident disease and complications of CKD.
References
 1.
Castledine C, Steenkamp R, Feest T, Tomson CR: UK renal registry 13th annual report (December 2010): chapter 6: survival and causes of death of UK adult patients on renal replacement therapy in 2009: national and centrespecific analyses. Nephron Clin Pract. 2011, 119 (Suppl 2): 10734.
 2.
Leoncini G, Viazzi F, Pontremoli R: Overall health assessment: a renal perspective. Lancet. 2010, 375 (9731): 20532054. 10.1016/S01406736(10)607489.
 3.
Roderick PJ, Raleigh VS, Hallam L, Mallick NP: The need and demand for renal replacement therapy in ethnic minorities in England. J Epidemiol Community Health. 1996, 50 (3): 334339. 10.1136/jech.50.3.334.
 4.
Gupta R: Burden of coronary heart disease in India. Indian Heart J. 2005, 57 (6): 632638.
 5.
Stevens L, Claybon MA, Schmid CH, Chen J, Horio M, Imai E, Nelson RG, Van Deventer M, Wang HY, Zuo L, Zhang YL, Levey AS: Evaluation of the chronic kidney disease epidemiology collaboration equation for estimating the glomerular filtration rate in multiple ethnicities. Kidney Int. 2011, 79: 555562. 10.1038/ki.2010.462.
 6.
Pottel H, Hoste L, Delanaye P, Cavalier E, Martens F: Demystifying ethnic/sex differences in kidney function: Is the difference in (estimating) glomerular filtration rate in serum creatinine concentration?. Clin Chim Acta. 2012, 413 (1920): 16121617. 10.1016/j.cca.2012.04.034.
 7.
Ortiz O, Russell M, Daley TL, Baumgartner RN, Waki M, Lichtman S, Wang J, Pierson RN, Heymsfield SB: Differences in skeletal muscle and bone mineral mass between black and white females and their relevance to estimates of body composition. Am J Clin Nutr. 1992, 55: 813.
 8.
Jones C, Jones CA, Wilson IB, Knox TA, Levey AS, Spiegelman D, Gorbach SL, Van Lente F, Stevens LA: Cystatin C and creatinine in an HIV cohort: the nutrition for healthy living study. Am J Kidney Dis. 2008, 51: 914924. 10.1053/j.ajkd.2008.01.027.
 9.
Gallagher D, Visser M, De Meersman RE, Sepúlveda D, Baumgartner RN, Pierson RN, Harris T, Heymsfield SB: Appendicular skeletal muscle mass: effects of age, gender, and ethnicity. J Appl Physiol. 1997, 83 (1): 229239.
 10.
Yajnik CS, Yudkin JS: The YY paradox. Lancet. 2004, 363 (9403): 16310.1016/S01406736(03)152695.
 11.
Rule AD, Bailey KR, Schwartz GL, Khosla S, Lieske JC, Melton LJ: For estimating creatinine clearance measuring muscle mass gives better results than those based on demographics. Kidney Int. 2009, 75 (10): 10711078. 10.1038/ki.2008.698.
 12.
Reddy K, Prabhakaran D, Chaturvedi V, Jeemon P, Thankappan KR, Ramakrishnan L, Mohan BV, Pandav CS, Ahmed FU, Joshi PP, Meera R, Amin RB, Ahuja RC, Das MS, Jaison TM: Methods for establishing a surveillance system for cardiovascular diseases in Indian industrial populations. Bull World Health Organ. 2006, 84: 461469. 10.2471/BLT.05.027037.
 13.
Lyngdoh T, Kinra S, Shlomo YB, Reddy S, Prabhakaran D, Smith GD, Ebrahim S: Indian Migration Study group. Sibrecruitment for studying migration and its impact on obesity and diabetes. Emerging Themes in Epidemiology. 2006, 3 (1): 210.1186/1742762232.
 14.
International Institute for Population Sciences (IIPS) National Family Health Survey (NFHS3). 2007, http://www.nfhsindia.org/nfhs3.html,
 15.
National Institute of Clinical Excellence CG127 Hypertension: The Clinical Management of Primary Hypertension in Adults. Aug 2011, National Clinical Guideline Centre, http://www.nice.org.uk/nicemedia/live/13561/56007/56007.pdf,
 16.
Wang ZM, Visser M, Ma R, Baumgartner RN, Kotler D, Gallagher D, Heymsfield SB: Skeletal muscle mass: evaluation of neutron activation and dualenergy Xray absorptiometry methods. J Appl Physiol. 1996, 80 (3): 824831.
 17.
Mosteller R: Simplified calculation of bodysurface area. N Engl J Med. 1987, 317: 1098
 18.
National Kidney Foundation: KDOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002, 39: S1266.
 19.
Levey AS, de Jong PE, Coresh J, El Nahas M, Astor BC, Matsushita K, Gansevoort RT, Kasiske BL, Eckardt KU: The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies Conference report. Kidney Int. 2011, 80 (1): 1728. 10.1038/ki.2010.483.
 20.
Ebrahim S, Kinra S, Bowen L, Andersen E, BenShlomo Y, Lyngdoh T, Ramakrishnan L, Ahuja RC, Joshi P, Das SM, Mohan M, Davey Smith G, Prabhakaran D, Reddy KS: Indian migration study group. The effect of ruraltourban migration on obesity and diabetes in india: A crosssectional study. PLoS Med. 2010, 7 (4): e100026810.1371/journal.pmed.1000268.
 21.
Hsu CY, Bansal N, Measured GFR as Gold Standard: All that glitters is not gold?. Clin J Am Soc Nephrol. 6 (8): 18131814.
 22.
Hsu C, Propert K, Xie D, Hamm L, He J, Miller E, Ojo A, Shlipak M, Teal V, Townsend R, Weir M, Wilson J, Feldman H, CRIC Investigators: Measured GFR does not outperform estimated GFR in predicting CKDrelated complications. J Am Soc Nephrol. 2011, 22 (10): 19311937. 10.1681/ASN.2010101077.
 23.
Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D: A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med. 1999, 130 (6): 461470.
 24.
Cockcroft D, Gault M: Prediction of creatinine clearance from serum creatinine. Nephron. 1976, 16 (1): 3141. 10.1159/000180580.
 25.
Kim J, Shen W, Gallagher D, Jones A, Wang Z, Wang J, Heshka S, Heymsfield SB: Totalbody skeletal muscle mass: estimation by a new dualenergy Xray absorptiometry method. Am J Clin Nutr. 2002, 76 (2): 378383.
 26.
Preiss D, Godber IM, Lamb EJ, Dalton RN, Gunn IR: The influence of a cookedmeat meal on estimated glomerular filtration rate. Ann Clin Biochem. 2007, 44: 3542. 10.1258/000456307779595995.
Prepublication history
The prepublication history for this paper can be accessed here:http://www.biomedcentral.com/14712369/14/30/prepub
Acknowledgements
The Indian Migration Study was funded by the Wellcome Trust (grant no.WT083707MA). Abstracts were presented at the ERAEDTA and UK Renal Association Conferences 2012. We would like to acknowledge the contribution made by Professor Andrew Rule, Associate Professor of Medicine, Mayo Clinic, USA.
The authors thank the field staff and local investigators who conducted this study.
The Indian Migration Study group comprises Prof. K Srinath Reddy, Dr. Dorairaj Prabhakaran, Prof. Tulsi Patel, Dr. Lakshmy Ramakrishnan, Dr. Ruby Gupta, and Dr. Tanica Lyngdoh (New Delhi); Prof. R C Ahuja and Prof. R K Saran (Lucknow); Dr. Prashant Joshi and Dr. N M Thakre (Nagpur); Dr. K V R Sarma, Prof. S Mohan Das, Dr. R K Jain, and Dr. S S Potnis (Hyderabad); Prof. Anura V Kurpad, Dr. Mario Vaz, AV Barathi, and Dr. Murali Mohan (Bangalore); Dr. Chittaranjan Yajnik (Pune); Prof. George Davey Smith and Prof. Yoav Ben Shlomo (Bristol); and Professor Shah Ebrahim and Dr. Sanjay Kinra (London School of Hygiene and Tropical Medicine).
Author information
Additional information
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SK SE YBS KR and HK contributed to study design. HK SK and SE managed the fieldwork, staff training and supervision of the data collection for the study. PB CT and YBS contributed to data analysis. PB CT YBS and DN contributed to data interpretation. PB wrote the first draft of the paper. CT YBS DN and HK contributed to the writing of the paper. All authors provided feedback on the manuscript and approved the submitted version.
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
Rights and permissions
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
About this article
Cite this article
Bailey, P.K., Tomson, C.R., Kinra, S. et al. Differences in estimation of creatinine generation between renal function estimating equations in an Indian population: crosssectional data from the Hyderabad arm of the Indian migration study. BMC Nephrol 14, 30 (2013) doi:10.1186/147123691430
Received
Accepted
Published
DOI
Keywords
 Creatinine
 Ethnicity
 Muscle mass
 Renal function