Objectives To perform a cross-sectional comparison of endogenous markers of glomerular filtration rate (GFR) (plasma symmetric dimethylarginine (SDMA) and estimated GFR (eGFR)) with a direct measure of GFR (using the plasma clearance of Inutest (In-GFR)), and a longitudinal evaluation of these markers in relation to the development of microalbuminuria, in young people with type 1 diabetes (T1D).
Methods Longitudinal stored blood samples (n=1105) were available from 417 young people from the Oxford Regional Prospective Study (an inception cohort of 527 children followed for a median of 10.3 (interquartile range 7.1–12.3) years), for measurement of SDMA and creatinine. Additional annually collected data on anthropometric parameters, HbA1c, insulin dose and three early morning albumin:creatinine ratios were available. In-GFR was measured in a representative subgroup of 183 subjects.
Main outcome measures SDMA and eGFR.
Results SDMA and eGFR were significantly and similarly associated with In-GFR (r=−0.38 and r=0.36, p<0.001). Overall SDMA levels were lower in microalbuminuric (n=116) than normoalbuminuric subjects (n=301) (0.385±0.063 vs 0.412±0.059 µmol/l, p<0.001), probably reflecting hyperfiltration. The pattern of change in SDMA levels with age differed between microalbuminuric and normoalbuminuric subjects. Whereas SDMA levels declined in both groups until the age of 16 years, thereafter they tended to rise only in microalbuminuric subjects, probably reflecting a decline in renal function.
Conclusions In this longitudinal study of young people with T1D, measurement of SDMA, in contrast to eGFR, proved to be a reliable marker in identifying changes in filtration rates associated with the development of microalbuminuria (MA).
Statistics from Altmetric.com
Diabetic nephropathy (DN) is a major determinant of morbidity and mortality among patients with type 1 diabetes (T1D).1 The earliest stage of clinical nephropathy is represented by microalbuminuria (MA) and its persistence is predictive of progression to overt nephropathy.1,–,3 The development of persistent M A is associated wit h increased arterial blood pressure and a progressive decline of renal function, with a high risk of developing end-stage renal disease and cardiovascular disease (CVD).1 Perturbations of glomerular filtration rate (GFR) are often seen during the early stages of DN, when there is a phase of hyperfiltration associated with hyperperfusion and increased renal size.4 5 Our group and others have reported that, during adolescence, glomerular hyperfiltration is an important predictor for the development of MA.2 6 7 In the large well-characterised cohort of subjects with T1D from the Oxford Regional Prospective Study (ORPS), recruited at diabetes diagnosis and followed through puberty, hyperfiltration was found to be a longitudinal determinant of MA independent of glycaemic control.6 Therefore, a careful long-term assessment of GFR in the T1D population may be important in assessing progression of renal disease.
What is already known on this topic
A longitudinal assessment of GFR in type 1 diabetes is highly recommended in order to detect variations in renal function and the associated risks.
A direct measurement of GFR, based on an intravenous infusion of an exogenous marker, is not easily applicable in a clinical setting.
Symmetric dimethylarginine (SDMA) is a reliable plasma marker of renal function in adults with chronic kidney disease.
What this study adds
In young people with type 1 diabetes, SDMA is independent of body size, and its levels are reduced in those developing microalbuminuria, probably reflecting hyperfiltration.
SDMA might represent a valuable plasma marker for monitoring changes in renal function in young people with type 1 diabetes and microalbuminuria.
Formal direct measurements of GFR are timeconsuming and require significant clinical and technical expertise to be performed accurately, particularly in hyperfiltration states.8 A simple and reliable plasma method to estimate/monitor GFR in patients with T1D would facilitate follow-up of subjects with incipient DN and the transition to overt nephropathy.9 In clinical practice, despite susceptibility to changes in production rate and renal tubular secretion, plasma creatinine is the most commonly used marker applied for estimating and monitoring GFR.10 11 However, the early detection of renal disease using plasma creatinine is confounded, particularly in childhood, by the high, betweenindividual, variability related to the effects of muscle mass and gender on production rates.12
Recently, it has been established that an endogenous marker, symmetric dimethylarginine (SDMA), related to the cardiovascular marker asymmetric dimethylarginine (ADMA),13 is a reliable estimate of GFR.14 A recent meta-analysis of all published data concluded that SDMA may indeed be a clinically valuable measure of GFR,14 being an endogenous product entirely removed from the body by renal excretion.14 In children, preliminary data suggest that plasma SDMA may be superior to plasma creatinine as an endogenous measure of GFR because it is independent of body size.15 The potential role of plasma SDMA in monitoring the early changes in GFR as well as its relationship with the development of MA in T1D patients has not been evaluated.
In the present study, we performed a cross-sectional comparison of plasma SDMA and eGFR with GFR measured using the plasma clearance of Inutest (In-GFR), together with a longitudinal evaluation of SDMA and eGFR in relation to the development of MA in young people with T1D.
Patients and methods
The ORPS cohort was established in 1986, and the characteristics of the cohort have been previously described.16 Children diagnosed under the age of 16 years, in the defined geographic region of Oxford Health Authority, were recruited within 3 months of diagnosis of diabetes to long-term follow-up. By the end of the recruitment period, in 1997, 527 subjects were included in the study. To date, the median follow-up is 10.3 years (interquartile range 7.1–12.3). Ethical approval was obtained from the district ethics committee in the region, with written consent from the parents. Children were asked to assent before entering the study.
Children were assessed at the end of the first year from diagnosis and annually thereafter. Annual assessment included measurements of height, weight, blood pressure and collection of three consecutive early morning urine samples for determination of the albumin-to-creatinine ratio (ACR, mg/mmol). Blood samples were collected for central laboratory measurement of HbA1c and additional samples were stored for each patient. Height was measured on wall-mounted stadiometers and weight measured on electronic scales. Body mass index (BMI) was calculated as the ratio between weight and the height squared and expressed in kg/m2.
Definition of microalbuminuria
MA was defined on the basis of an ACR between 3.5 and 35 mg/ mmol in males and between 4.0 and 47 mg/mmol in females in two out of three consecutive early-morning urine collections during an annual assessment.16 Persistent MA was defined as an ACR within the microalbuminuric range based on two out three urines or two out of two urines each year for at least two consecutive years. Transient MA was defined as the presence of MA for a single year with subsequent regression to normal.
In a representative subgroup of 183 patients, a direct measurement of GFR (In-GFR) was made at 5 years’ diabetes duration, by plasma clearance of Inutest (Laevosan-Gesellschaft, Linz, Austria), a branched chain polyfructosan with a Stokes radius profile equivalent to inulin.6 The test was performed by using a single intravenous bolus and numerical analysis17 of the concentration–time curve.18
Longitudinally stored plasma samples (n=1105) from a larger group of 417 children, adolescents and young adults (239 male/178 female) (median samples 3.5, range 1–6 from each patient) were available for the measurement of plasma SDMA and creatinine.
Additional data on height, weight, BMI, HbA1c, insulin dose (U/kg/day) and diabetes duration at the time of blood collection were available for all the study population.
Plasma SDMA and creatinine were measured by stable isotope dilution electrospray mass spectrometry-mass spectrometry (MSMS) as previously described,19 and results were calculated using Analyst version 1.4.1. The intra-assay variation for SDMA (mean CV%, n=10) was 0.587 μmol/l, 2.1%, and the interassay variation (n=57) was as follows: 0.409 μmol/l, 3.2%; 1.884 μmol/l, 2.4%; 4.412 μmol/l, 3.3%. The interassay variation for creatinine (n=56) was as follows: 64.8 μmol/l, 3.5%; 183.8 μmol/l, 2.4%; 495.5 μmol/l, 3.1%.
Urinary albumin was measured with an ELISA method. The within- and between-assay CVs was 6% and 12%, respectively.
Urinary creatinine was measured using a modified Jaffe method (Unimate 7; Roche Diagnostic Systems, Switzerland) on a Cobas Mira (Roche Diagnostic Systems) automated spectrophotometer. The CV was 2% at 2.2 mmol/l.
HbA1c was measured in a central laboratory, initially by an electrophoretic method (Ciba Corning Diagnostics, Halstead, UK) and then by high-performance liquid chromatography (HPLC-DIAMAT; Bio-Rad, Hemel Hempstead, UK). The relationship between the two methods was carefully evaluated and has been described previously.16
Inutest was analysed as fructose following preincubation of the samples to remove free fructose and acid hydrolysis of the Inutest to yield fructose as previously described.6
For subjects ≤18 years, eGFR was calculated as follows: 42×height (cm)/creatinine (μmol/l), where 42 is the local laboratory-derived constant.20
For subjects >18 years, eGFR was calculated by the Modification of Diet in Renal Disease (MDRD) formula: (175×(creatinine (μmol/l)/88.4)−1.154×(age)−0.203×(0.742 if patient is female)×(1.201 if patient is black).21
Data were analysed for normality using the Kolmorgorov– Smirnov test, and log-transformed to normal distributions wherever necessary to allow use of parametric tests. Results are reported as means±SD or median(range) unless otherwise stated. An unpaired t test was used to assess differences between microalbuminuric and normoalbuminuric subjects. Cross-sectional associations between SDMA or eGFR and other variables of interest were performed by univariate and multiple regression analysis, and the results are expressed as B coefficient (95% CI) and correlation coefficients (r). In addition, a linear mixed effect model was used to analyse the data longitudinally. General linear models were used to test interactions of the relationship between SDMA and age, adjusting for potential confounders, with MA status.
All analyses were performed using SPSS for Windows version 11.5 (SPSS, Chicago), apart from the linear mixed model, which was performed with S-plus software. Multilevel modelling (MLwiN; Institute of Education, London) was also used to further examine longitudinal differences in SDMA between microalbuminuric and normoalbuminuric subjects.22 p<0.05 was taken as significant for all analyses.
Association of SDMA and eGFR with In-GFR
In 183 subjects, a direct measurement of GFR (In-GFR) was performed using plasma clearance from Inutest. The clinical and biochemical characteristics of this group are shown in table 1.
This subgroup of 183 subjects, did not differ in age, height, BMI and metabolic control from the larger cohort.
A modest and inverse correlation was found between plasma SDMA and In-GFR (r=−0.38, p<0.001) (figure 1A). This association was strengthened (r=−0.42, p<0.001) after adjusting for covariates, including age (r=−0.21, p=0.005), BMI (r=−0.18; p=0.02), HbA1c (r=−0.25, p=0.001), gender (p=0.005; higher levels in male) and insulin dose (r=−0.13, p=0.09). In a multiple regression model, in addition to In-GFR, only HbA1c and gender were independently related to SDMA (table 2). A significant correlation was also found between eGFR and In-GFR (r=0.36, p<0.001) (figure 1B).
Longitudinal evaluation of SDMA
The entire study population consisted of 239 males and 178 females. The subjects anthropometric and biochemical characteristics are shown in table 1. There were no significant differences between males and females for age, T1D duration and HbA1c. Conversely, BMI was significantly higher in females than in males (p=0.005). Significant differences in SDMA and eGFR were obser ved bet ween genders, wit h males having higher SDMA levels (0.420±0.060 vs 0.385±0.060 μmol/litre, p<0.001) and lower eGFR (137±25 vs 144±28 ml/min/1.73 m2, p=0.008), and these differences persisted after adjustment for BMI.
In this larger cohort, factors significantly related to changes in SDMA during the follow-up period, in the univariate analysis, were gender (low levels in female), BMI, HbA1c and insulin dose (table 3). However, in a multiple-regression model (including all variables significantly related to SDMA in the univariate analysis plus age) the only factors independently related to SDMA were gender, HbA1c and insulin dose (table 3). Factors related to eGFR in the univariate analysis were age, duration, BMI, height, gender and HbA1c (table 3). In the multiple- regression model age, gender and HbA1c remained significantly related to eGFR (table 3).
SDMA and eGFR in subjects with and without microalbuminuria
Significant differences were found between subjects with MA (n=116; 62 with persistent MA; 54 with transient MA) and normoalbuminuric subjects (n=301) in SDMA, with lower values in the MA group (0.385±0.063 vs 0.412±0.059 μmol/litre, p<0.001), and higher values of eGFR (144.7±30.7 vs 137.8±24.3 ml /min /1.73 m2, p=0.03). Patients with MA were older (16.2±4.9 vs 15.2±4.4 years, p=0.03), had a longer duration of diabetes (7.3±3.6 vs 6.2±3.4 years, p=0.005) and had higher BMI (22±5 vs 21±4 kg/ m2) and HbA1c (10.5±2.0 vs 9.6±1.6%, p<0.001). After controlling for these variables and gender, differences between the two groups persisted for SDMA (p=0.02) but not for eGFR (p=0.08).
The longitudinal evaluation of the pattern of change in SDMA levels with age demonstrated a significant difference between subjects with and without MA (figure 2A). Whereas levels of SDMA declined in both groups until the age of 16 years, thereafter they tended to rise only in subjects with MA but not in those with normoalbuminuria. Age-related differences in SDMA between the two groups persisted after adjusting for diabetes duration, gender, BMI, HbA1c and insulin dose (adjusted p for interaction=0.03). In addition, within the MA group, the age-related variations in SDMA were more marked in those with persistent MA than in those with transient MA (figure 2A). In contrast, no significant differences were detected in the pattern of change in eGFR with age between MA positive and normoalbuminuric subjects (figure 2B). Plasma SDMA levels showed relevant changes in relation to years to MA onset. In particular, there was a progressive decline in SDMA before MA onset, followed by a gradual increase in both subjects with transient and persistent MA (figure 3).
In the present study, we explored SDMA and eGFR as possible measures of renal function in a large cohort of young people with childhood-onset T1D followed longitudinally from diagnosis. As in other populations, including adults with T1D,14 23 we found that SDMA and eGFR were significantly correlated with the direct measurement of GFR using Inutest. SDMA in contrast to creatinine-derived eGFR was independent of body size and was more sensitive in tracking changes in renal function, suggesting that this endogenous biomarker of GFR could be superior to creatinine-based eGFR in assessing renal function during childhood and adolescence in T1D subjects.
SDMA, like its stereoisomer ADMA, is a dimethylarginine produced endogenously by post-translational methylation of arginine residues in proteins and is released within the cells and then into the plasma during proteolysis.24 SDMA is produced by all nucleated cells and is entirely eliminated by renal excretion.25 26 Consequently, previous studies, conducted mainly in patients with chronic kidney disease, have established that SDMA has some of the properties of a reliable plasma biomarker of GFR, including excellent correlation with GFR assessed by clearance methods.14 In contrast, the almost universal use of plasma creatinine as an endogenous marker of renal function, even after correction for age, gender and body composition, is still limited by poor analytical methodology,27 diet28 and changes in tubular secretion as GFR declines.29
The present study demonstrates a relatively weak, but significant, correlation between plasma SDMA and In-GFR in children, adolescents, and young adults with T1D. Previous studies of subjects with a wide range of GFR demonstrated a high correlation between SDMA and GFR by direct measurement (r=0.8–0.9);14 however those studies rarely included patients with hyperfiltration. Measuring and monitoring GFR in the hyperfiltration range are difficult, as clearly shown in the cohort where the MDRD formula for GFR calculation was validated.21 In fact, even though the MDRD formula performed well in the whole cohort, in subjects with GFR above 80 ml/min/1.73 m2 there was a large variability in the results.21 Therefore, a possible explanation for the weaker association between SDMA and In-GFR in our study could be the narrower range of GFR and the inclusion of hyperfiltering subjects.
In this cohort, plasma SDMA was independent of height, and this finding is consistent with previous data in non-diabetic paediatric renal failure patients.15 Since the calculation of GFR with formulae based on plasma creatinine in children and adolescents requires accurate measurement of height, plasma SDMA might provide a practical alternative for assessing GFR in this age group. Gender-related differences in SDMA levels emerged in this study, and this finding might be related to variations in SDMA production30 or increased catabolism.31 However, as the risk of MA, and possibly hyperfiltration, is higher in adolescent girls,16 the lower levels in girls might be related to a higher GFR. A limitation for the interpretation of this finding is the lack of data on normal children. Consequently, it is difficult to completely discriminate changes in SDMA related to diabetes from normal variations due to growth and puberty. However, these gender-related differences can be taken into account when applying SDMA as a measure of renal function in individuals with diabetes.
Interestingly, there was a significant inverse association of plasma SDMA with HbA1c. This suggests, based also on the direct relationship found between In-GFR and HbA1c (data not shown), that persistent hyperglycaemia might be a significant driver of hyperfiltration. However, an effect of chronic hyperglycaemia on dimethylarginine synthesis and/or release cannot be excluded. A negative association between insulin dose and SDMA was also detected in this cohort. This might be related to an inhibitory effect of insulin on proteolysis and consequently to a reduced release of SDMA from cells or to an insulin-mediated downregulation of the enzymes implicated in the methylation of arginine residues in proteins.24
Despite these limitations, the study uncovered significant age-related differences in plasma SDMA, between subjects with and without MA. Overall, plasma SDMA was significantly lower in subjects developing MA, and this might be related to greater hyperfiltration. In addition, in subjects developing MA, particularly in those with persistent MA, the longitudinal data revealed a decline in SDMA during puberty, until about 16 years of age, followed by an increase. In contrast, in normoalbuminuric subjects there was a slightly progressive decline in SDMA with age, similar to that recently documented for ADMA in healthy children and adolescents.32 These data are consistent with a hyperfiltration stage in MA subjects leading to progressive decline in renal function. The similar pattern of SDMA changes in patients with transient and persistent MA, underlines the concept that even transient MA is expression of renal pathology and, as recently shown by our group, is an independent risk factor for progression to more advanced stages of nephropathy.33 Interestingly, in this ORPS cohort, the median age for the development of MA is around 16 years,33 and a previous investigation has shown a clear role of hyperfiltration in predicting the subsequent development of MA, independently of glycaemic control,6 and a progressive decline in renal function after the onset of MA.6
When we assessed SDMA in relation to the time of MA onset, there was a progressive decline before the development of MA followed by a gradual increase, in both subjects with transient and persistent MA, and this finding supports our hypothesis that the U-shaped pattern in SDMA in relation to age might be due to variations in renal function associated with the development of MA. Taken together, these observations suggest that plasma SDMA might be a relatively simple, non-invasive, biomarker for monitoring variations in renal function in T1D subjects.
In contrast, eGFR did not show any significant difference in the pattern of change with age between MA and normoalbuminuric subjects. This highlights a major problem in using creatinine-based eGFR measures to monitor GFR, particularly in the childhood–adolescence–adult transition, when a shift from a paediatric creatinine-based formula, taking in account height9 to the recommended MDRD formula21 in adults is required. Thus, even though eGFR correlated well with the direct measurement of GFR, it proved to be less sensitive than plasma SDMA in tracking changes in renal function. Furthermore, since the calculation of eGFR in children requires accurate measurement of height, plasma SDMA might provide a practical alternative for assessment of GFR in the paediatric age group.
A factor limiting the general application of plasma SDMA, for routine monitoring of GFR, is the need to access highly sensitive MSMS. This is not a consideration for major cohort studies, even though it is only available in relatively few clinical laboratories. However, as a result of the routine clinical application of MSMS for screening and monitoring of a wide variety of conditions, including immunosuppressive therapy, MSMS analysis is more widely available than might be expected. In addition, the cost of measuring plasma SDMA, including instrument amortisation, is significantly less than the cost of performing a direct GFR measurement.34 Furthermore, measuring SDMA might represent more than a simple assessment of GFR, as recent studies have emphasised the relationship between SDMA and cardiovascular risk in patients with chronic renal disease35 as well as in the general population.36
In conclusion, in this longitudinal study of T1D, measurement of SDMA proved to be a sensitive biomarker, independent of body size, in tracking changes in renal function associated with the development of MA. Serial measurements of SDMA may prove a valuable plasma marker for monitoring the development of hyperfiltration and the transition to a progressive decline in renal function.
We acknowledge the Juvenile Diabetes Research Foundation, the National Institute for Health Research (NIHR) Cambridge Comprehensive Biomedical Research Centre, the study field workers, the laboratory assistance of Angie Watts and Dot Harris, the Barts-Oxford Study field workers, paediatricians, physicians and diabetes nurse specialists in the Oxford Region.
Members of the Oxford Regional Prospective Study Steering Committee are as follows: DB Dunger, RN Dalton, J Fuller, EAM Gale, H Keen, M Murphy, HAW Niel, RJ Young and T Konopelska-Bahu. Members of the Oxford Regional Prospective Study are as follows: J Edge, John Radcliffe Hospital Oxford; HAW Neil and D Matthews, The Oxford Centre for Diabetes, Endocrinology and Metabolism, The Churchill Hospital, Oxford; RAF Bell and A Taylor, Horton General Hospital, Banbury; A Mukhtar, BP O’Malley, BR Silk and EH Smith, Kettering District Hospital, Kettering; RDM Scott, King Edward VII Hospital; FM Ackland, CJ Fox and NK Griffin, Northampton General Hospital; N Mann, H Simpson, P Cove Smith and M Pollitzer, Royal Berkshire Hospital, Reading; RS Brown and AH Knight, Stoke Mandeville Hospital, Aylesbury; JM Cowen and JC Pearce, Wexham Park Hospital, Slough.
Competing interests None.
Ethics approval Ethics approval was provided by the district ethics committee in the Oxford region.
Provenance and peer review Not commissioned; externally peer reviewed.
Funding The Oxford Regional Prospective Study is funded by Diabetes UK. In addition, this study was supported (in part) by a European Society for Pediatric Endocrinology Research Fellowship, sponsored by Novo Nordisk A/S (to MLM).
Patient consent Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.